ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1053 - (show annotations)
Tue Oct 9 15:43:08 2012 UTC (8 years, 11 months ago) by zherczeg
File size: 237858 byte(s)
Improved CMake support under Windows.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.32
5 ------------
7 1. Improved JIT compiler optimizations for first character search and single
8 character iterators.
10 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
11 Patch by Daniel Richard G.
13 3. Single character iterator optimizations in the JIT compiler.
15 4. Improved JIT compiler optimizations for character ranges.
17 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
18 Reported by Giuseppe D'Angelo.
20 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
21 of a line, was being set incorrectly in cases where .* appeared inside
22 atomic brackets at the start of a pattern, or where there was a subsequent
23 *PRUNE or *SKIP.
25 7. Improved instruction cache flush for POWER/PowerPC.
26 Patch by Daniel Richard G.
28 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
29 grep:
31 (a) There is now no limit to the number of patterns to be matched.
33 (b) An error is given if a pattern is too long.
35 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
36 are now supported.
38 (d) --exclude-from and --include-from (multiple use) have been added.
40 (e) Exclusions and inclusions now apply to all files and directories, not
41 just to those obtained from scanning a directory recursively.
43 (f) Multiple uses of -f and --file-list are now supported.
45 (g) In a Windows environment, the default for -d has been changed from
46 "read" (the GNU grep default) to "skip", because otherwise the presence
47 of a directory in the file list provokes an error.
49 (h) The documentation has been revised and clarified in places.
51 9. Improve the matching speed of capturing brackets.
53 10. Changed the meaning of \X so that it now matches a Unicode extended
54 grapheme cluster.
56 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
57 out POSIX threads when JIT support is configured.
59 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
61 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
62 structure could go wrong in environments where size_t is not the same size
63 as int.
65 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
66 set.
68 15. The EBCDIC support had decayed; later updates to the code had included
69 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
70 general tidy up of EBCDIC-related issues, and the documentation was also
71 not quite right. There is now a test that can be run on ASCII systems to
72 check some of the EBCDIC-related things (but is it not a full test).
74 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
75 in a small tidy to the code.
77 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
79 18. If the --only-matching (-o) option in pcregrep is specified multiple
80 times, each one causes appropriate output. For example, -o1 -o2 outputs the
81 substrings matched by the 1st and 2nd capturing parentheses. A separating
82 string can be specified by --om-separator (default empty).
84 19. Improving the first n character searches.
86 20. Turn case lists for horizontal and vertical white space into macros so that
87 they are defined only once.
89 21. This set of changes together give more compatible Unicode case-folding
90 behaviour for characters that have more than one other case when UCP
91 support is available.
93 (a) The Unicode property table now has offsets into a new table of sets of
94 three or more characters that are case-equivalent. The MultiStage2.py
95 script that generates these tables (the pcre_ucd.c file) now scans
96 CaseFolding.txt instead of UnicodeData.txt for character case
97 information.
99 (b) The code for adding characters or ranges of characters to a character
100 class has been abstracted into a generalized function that also handles
101 case-independence. In UTF-mode with UCP support, this uses the new data
102 to handle characters with more than one other case.
104 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
105 whose other case is greater than 256 are now correctly matched
106 caselessly. Previously, the high codepoint matched the low one, but not
107 vice versa.
109 (d) The processing of \h, \H, \v, and \ in character classes now makes use
110 of the new class addition function, using character lists defined as
111 macros alongside the case definitions of 20 above.
113 (e) Caseless back references now work with characters that have more than
114 one other case.
116 (f) General caseless matching of characters with more than one other case
117 is supported.
119 22. Unicode character properties were updated from Unicode 6.2.0
121 23. Improved CMake support under Windows. Patch by Daniel Richard G.
124 Version 8.31 06-July-2012
125 -------------------------
127 1. Fixing a wrong JIT test case and some compiler warnings.
129 2. Removed a bashism from the RunTest script.
131 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
132 to unsigned type, result still unsigned" that was given by an MS compiler
133 on encountering the code "-sizeof(xxx)".
135 4. Partial matching support is added to the JIT compiler.
137 5. Fixed several bugs concerned with partial matching of items that consist
138 of more than one character:
140 (a) /^(..)\1/ did not partially match "aba" because checking references was
141 done on an "all or nothing" basis. This also applied to repeated
142 references.
144 (b) \R did not give a hard partial match if \r was found at the end of the
145 subject.
147 (c) \X did not give a hard partial match after matching one or more
148 characters at the end of the subject.
150 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
151 a partial match for the string "\r".
153 (e) When newline was set to CRLF, the metacharacter "." did not recognize
154 a partial match for a CR character at the end of the subject string.
156 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
157 running pcretest, the text "(JIT)" added to the output whenever JIT is
158 actually used to run the match.
160 7. Individual JIT compile options can be set in pcretest by following -s+[+]
161 or /S+[+] with a digit between 1 and 7.
163 8. OP_NOT now supports any UTF character not just single-byte ones.
165 9. (*MARK) control verb is now supported by the JIT compiler.
167 10. The command "./RunTest list" lists the available tests without actually
168 running any of them. (Because I keep forgetting what they all are.)
172 12. Applied a (slightly modified) user-supplied patch that improves performance
173 when the heap is used for recursion (compiled with --disable-stack-for-
174 recursion). Instead of malloc and free for each heap frame each time a
175 logical recursion happens, frames are retained on a chain and re-used where
176 possible. This sometimes gives as much as 30% improvement.
178 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
179 call.
181 14. As documented, (*COMMIT) is now confined to within a positive assertion.
183 15. It is now possible to link pcretest with libedit as an alternative to
184 libreadline.
186 16. (*COMMIT) control verb is now supported by the JIT compiler.
188 17. The Unicode data tables have been updated to Unicode 6.1.0.
190 18. Added --file-list option to pcregrep.
192 19. Added binary file support to pcregrep, including the -a, --binary-files,
193 -I, and --text options.
195 20. The madvise function is renamed for posix_madvise for QNX compatibility
196 reasons. Fixed by Giuseppe D'Angelo.
198 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
200 22. Changed the option for creating symbolic links for 16-bit man pages from
201 -s to -sf so that re-installing does not cause issues.
203 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
205 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
206 matches in certain environments (the workspace was not being correctly
207 retained). Also added to pcre_dfa_exec() a simple plausibility check on
208 some of the workspace data at the beginning of a restart.
210 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
211 was not doing so when it should - probably a typo introduced by SVN 528
212 (change 8.10/14).
214 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
215 \w+ when the character tables indicated that \x{c4} was a word character.
216 There were several related cases, all because the tests for doing a table
217 lookup were testing for characters less than 127 instead of 255.
219 27. If a pattern contains capturing parentheses that are not used in a match,
220 their slots in the ovector are set to -1. For those that are higher than
221 any matched groups, this happens at the end of processing. In the case when
222 there were back references that the ovector was too small to contain
223 (causing temporary malloc'd memory to be used during matching), and the
224 highest capturing number was not used, memory off the end of the ovector
225 was incorrectly being set to -1. (It was using the size of the temporary
226 memory instead of the true size.)
228 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
229 ovector size, it uses memory at the end of the block that it has got.
231 29. Check for an overlong MARK name and give an error at compile time. The
232 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
234 30. JIT compiler update.
236 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
237 Rill for the patch.
239 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
241 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
243 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
244 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
245 was enabled.
247 35. Improve JIT code generation for greedy plus quantifier.
249 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
250 1 to "aa" instead of to an empty string. The bug affected repeated groups
251 that could potentially match an empty string.
253 37. Optimizing single character iterators in JIT.
255 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
256 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
257 codepoints that are too big for the mode are faulted, and in a UTF mode,
258 disallowed codepoints are also faulted.
260 39. If PCRE was compiled with UTF support, in three places in the DFA
261 matcher there was code that should only have been obeyed in UTF mode, but
262 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
263 processing when bytes with values greater than 127 were present. In 16-bit
264 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
265 both cases the values are those that cannot be the first data item in a UTF
266 character. The three items that might have provoked this were recursions,
267 possessively repeated groups, and atomic groups.
269 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
270 and pcregrep, because some OS require shared objects to be explicitly
271 passed to ld, causing the link step to fail if they are not.
273 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
274 mode, patterns that started with \h* or \R* might be incorrectly matched.
277 Version 8.30 04-February-2012
278 -----------------------------
280 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
281 name is defined in ctype.h.
283 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
284 only in quite long subpatterns.
286 3. Removed the function pcre_info(), which has been obsolete and deprecated
287 since it was replaced by pcre_fullinfo() in February 2000.
289 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
290 match a (*MARK), and the match failed at the start of the subject, a
291 reference to memory before the start of the subject could occur. This bug
292 was introduced by fix 17 of release 8.21.
294 5. A reference to an unset group with zero minimum repetition was giving
295 totally wrong answers (in non-JavaScript-compatibility mode). For example,
296 /(another)?(\1?)test/ matched against "hello world test". This bug was
297 introduced in release 8.13.
299 6. Add support for 16-bit character strings (a large amount of work involving
300 many changes and refactorings).
302 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
303 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
304 from a file.
306 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
307 rounding is not applied in this particular case).
309 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
310 if they appear, or are escaped, in patterns.
312 10. Get rid of a number of -Wunused-but-set-variable warnings.
314 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
315 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
316 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
317 also returns the mark "x". This bug applied to capturing parentheses,
318 non-capturing parentheses, and atomic parentheses. It also applied to some
319 assertions.
321 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
322 information out of configure.ac instead of relying on pcre.h.generic, which
323 is not stored in the repository.
325 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
326 -lreadline.
328 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
330 15. Applied Graycode's patch to put the top-level frame on the stack rather
331 than the heap when not using the stack for recursion. This gives a
332 performance improvement in many cases when recursion is not deep.
334 16. Experimental code added to "pcretest -C" to output the stack frame size.
337 Version 8.21 12-Dec-2011
338 ------------------------
340 1. Updating the JIT compiler.
342 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
343 are added as well.
345 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
346 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
347 calling _pcre_jit_exec. Some extra comments are added.
349 4. (*MARK) settings inside atomic groups that do not contain any capturing
350 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
351 was introduced by change 18 for 8.20.
353 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
354 ECMA-262 standard.
356 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
357 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
358 This bug was probably introduced by change 9 of 8.13.
360 7. While fixing 6 above, I noticed that a number of other items were being
361 incorrectly rejected as "not fixed length". This arose partly because newer
362 opcodes had not been added to the fixed-length checking code. I have (a)
363 corrected the bug and added tests for these items, and (b) arranged for an
364 error to occur if an unknown opcode is encountered while checking for fixed
365 length instead of just assuming "not fixed length". The items that were
366 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
367 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
368 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
370 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
371 being incorrectly compiled and would have given unpredicatble results.
373 9. A possessively repeated subpattern with minimum repeat count greater than
374 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
375 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
376 the first (A) could occur when it should not.
378 10. Add a cast and remove a redundant test from the code.
380 11. JIT should use pcre_malloc/pcre_free for allocation.
382 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
383 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
384 is anything other than /usr, -L is still shown).
386 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
388 14. Perl does not support \N without a following name in a [] class; PCRE now
389 also gives an error.
391 15. If a forward reference was repeated with an upper limit of around 2000,
392 it caused the error "internal error: overran compiling workspace". The
393 maximum number of forward references (including repeats) was limited by the
394 internal workspace, and dependent on the LINK_SIZE. The code has been
395 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
396 the default depends on LINK_SIZE. There is a new upper limit (for safety)
397 of around 200,000 forward references. While doing this, I also speeded up
398 the filling in of repeated forward references.
400 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
401 incorrectly expecting the subject to contain another "a" after the start.
403 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
404 in the match, the SKIP should be ignored. This was not happening; instead
405 the SKIP was being treated as NOMATCH. For patterns such as
406 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
407 tested.
409 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
410 now much more compatible with Perl, in particular in cases where the result
411 is a non-match for a non-anchored pattern. For example, if
412 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
413 "m", where previously it did not return a name. A side effect of this
414 change is that for partial matches, the last encountered mark name is
415 returned, as for non matches. A number of tests that were previously not
416 Perl-compatible have been moved into the Perl-compatible test files. The
417 refactoring has had the pleasing side effect of removing one argument from
418 the match() function, thus reducing its stack requirements.
420 19. If the /S+ option was used in pcretest to study a pattern using JIT,
421 subsequent uses of /S (without +) incorrectly behaved like /S+.
423 21. Retrieve executable code size support for the JIT compiler and fixing
424 some warnings.
426 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
427 not work when the shorter character appeared right at the end of the
428 subject string.
430 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
431 systems.
433 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
434 output it when the /M option is used in pcretest.
436 25. The CheckMan script was not being included in the distribution. Also, added
437 an explicit "perl" to run Perl scripts from the PrepareRelease script
438 because this is reportedly needed in Windows.
440 26. If study data was being save in a file and studying had not found a set of
441 "starts with" bytes for the pattern, the data written to the file (though
442 never used) was taken from uninitialized memory and so caused valgrind to
443 complain.
445 27. Updated RunTest.bat as provided by Sheri Pierce.
447 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
449 29. Computation of memory usage for the table of capturing group names was
450 giving an unnecessarily large value.
453 Version 8.20 21-Oct-2011
454 ------------------------
456 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
457 a POSIX class. After further experiments with Perl, which convinced me that
458 Perl has bugs and confusions, a closing square bracket is no longer allowed
459 in a POSIX name. This bug also affected patterns with classes that started
460 with full stops.
462 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
463 captured substring, but while checking the failing first alternative,
464 substring 1 is temporarily captured. If the output vector supplied to
465 pcre_exec() was not big enough for this capture, the yield of the function
466 was still zero ("insufficient space for captured substrings"). This cannot
467 be totally fixed without adding another stack variable, which seems a lot
468 of expense for a edge case. However, I have improved the situation in cases
469 such as /(a)(b)x|abc/ matched against "abc", where the return code
470 indicates that fewer than the maximum number of slots in the ovector have
471 been set.
473 3. Related to (2) above: when there are more back references in a pattern than
474 slots in the output vector, pcre_exec() uses temporary memory during
475 matching, and copies in the captures as far as possible afterwards. It was
476 using the entire output vector, but this conflicts with the specification
477 that only 2/3 is used for passing back captured substrings. Now it uses
478 only the first 2/3, for compatibility. This is, of course, another edge
479 case.
481 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
482 main code base, and can be used by building with --enable-jit. When this is
483 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
484 runtime --no-jit option is given.
486 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
487 ovector, the return from the function was zero, implying that there were
488 other matches that did not fit. The correct "exactly full" value is now
489 returned.
491 6. If a subpattern that was called recursively or as a subroutine contained
492 (*PRUNE) or any other control that caused it to give a non-standard return,
493 invalid errors such as "Error -26 (nested recursion at the same subject
494 position)" or even infinite loops could occur.
496 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
497 computing the minimum length on reaching *ACCEPT, and so ended up with the
498 wrong value of 1 rather than 0. Further investigation indicates that
499 computing a minimum subject length in the presence of *ACCEPT is difficult
500 (think back references, subroutine calls), and so I have changed the code
501 so that no minimum is registered for a pattern that contains *ACCEPT.
503 8. If (*THEN) was present in the first (true) branch of a conditional group,
504 it was not handled as intended. [But see 16 below.]
506 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
507 Sheri Pierce.
509 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
510 the first byte in a match must be "a".
512 11. Change 17 for 8.13 increased the recursion depth for patterns like
513 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
514 pattern contains any instances of (*THEN). If it does not, the old
515 optimizations are restored. It would be nice to do this on a per-group
516 basis, but at the moment that is not feasible.
518 12. In some environments, the output of pcretest -C is CRLF terminated. This
519 broke RunTest's code that checks for the link size. A single white space
520 character after the value is now allowed for.
522 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
523 For "fr", it uses the Windows-specific input and output files.
525 14. If (*THEN) appeared in a group that was called recursively or as a
526 subroutine, it did not work as intended. [But see next item.]
528 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
529 pattern fragments (but not containing any | characters). If A and B are
530 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
531 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
532 D. In other words, Perl considers parentheses that do not contain any |
533 characters to be part of a surrounding alternative, whereas PCRE was
534 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
535 differently. PCRE now behaves in the same way as Perl, except in the case
536 of subroutine/recursion calls such as (?1) which have in any case always
537 been different (but PCRE had them first :-).
539 16. Related to 15 above: Perl does not treat the | in a conditional group as
540 creating alternatives. Such a group is treated in the same way as an
541 ordinary group without any | characters when processing (*THEN). PCRE has
542 been changed to match Perl's behaviour.
544 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
545 RunGrepTest script failed.
547 18. Change 22 for version 13 caused atomic groups to use more stack. This is
548 inevitable for groups that contain captures, but it can lead to a lot of
549 stack use in large patterns. The old behaviour has been restored for atomic
550 groups that do not contain any capturing parentheses.
552 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
553 suppress the check for a minimum subject length at run time. (If it was
554 given to pcre_exec() or pcre_dfa_exec() it did work.)
556 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
557 fail to work when decoding hex characters in data strings in EBCDIC
558 environments.
560 21. It appears that in at least one Mac OS environment, the isxdigit() function
561 is implemented as a macro that evaluates to its argument more than once,
562 contravening the C 90 Standard (I haven't checked a later standard). There
563 was an instance in pcretest which caused it to go wrong when processing
564 \x{...} escapes in subject strings. The has been rewritten to avoid using
565 things like p++ in the argument of isxdigit().
568 Version 8.13 16-Aug-2011
569 ------------------------
571 1. The Unicode data tables have been updated to Unicode 6.0.0.
573 2. Two minor typos in pcre_internal.h have been fixed.
575 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
576 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
577 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
579 4. There were a number of related bugs in the code for matching backrefences
580 caselessly in UTF-8 mode when codes for the characters concerned were
581 different numbers of bytes. For example, U+023A and U+2C65 are an upper
582 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
583 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
584 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
585 2-byte code at the end of the subject (it thought there wasn't enough data
586 left).
588 5. Comprehensive information about what went wrong is now returned by
589 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
590 as the output vector has at least 2 elements. The offset of the start of
591 the failing character and a reason code are placed in the vector.
593 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
594 now returned is for the first byte of the failing character, instead of the
595 last byte inspected. This is an incompatible change, but I hope it is small
596 enough not to be a problem. It makes the returned offset consistent with
597 pcre_exec() and pcre_dfa_exec().
599 7. pcretest now gives a text phrase as well as the error number when
600 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
601 failure, the offset and reason code are output.
603 8. When \R was used with a maximizing quantifier it failed to skip backwards
604 over a \r\n pair if the subsequent match failed. Instead, it just skipped
605 back over a single character (\n). This seems wrong (because it treated the
606 two characters as a single entity when going forwards), conflicts with the
607 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
608 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
609 has been changed.
611 9. Some internal refactoring has changed the processing so that the handling
612 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
613 time (the PCRE_DOTALL option was changed this way some time ago: version
614 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
615 which was always a bit of a fudge. It also means that there is one less
616 argument for the match() function, which reduces its stack requirements
617 slightly. This change also fixes an incompatibility with Perl: the pattern
618 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
620 10. More internal refactoring has drastically reduced the number of recursive
621 calls to match() for possessively repeated groups such as (abc)++ when
622 using pcre_exec().
624 11. While implementing 10, a number of bugs in the handling of groups were
625 discovered and fixed:
627 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
628 (a|)*(?1) gave a compile-time internal error.
629 ((a|)+)+ did not notice that the outer group could match an empty string.
630 (^a|^)+ was not marked as anchored.
631 (.*a|.*)+ was not marked as matching at start or after a newline.
633 12. Yet more internal refactoring has removed another argument from the match()
634 function. Special calls to this function are now indicated by setting a
635 value in a variable in the "match data" data block.
637 13. Be more explicit in pcre_study() instead of relying on "default" for
638 opcodes that mean there is no starting character; this means that when new
639 ones are added and accidentally left out of pcre_study(), testing should
640 pick them up.
642 14. The -s option of pcretest has been documented for ages as being an old
643 synonym of -m (show memory usage). I have changed it to mean "force study
644 for every regex", that is, assume /S for every regex. This is similar to -i
645 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
646 using it. It makes it easier to run collections of tests with and without
647 study enabled, and thereby test pcre_study() more easily. All the standard
648 tests are now run with and without -s (but some patterns can be marked as
649 "never study" - see 20 below).
651 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
652 restoration of the capturing data to the outer values was not happening
653 correctly.
655 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
656 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
657 pattern had matched an empty string, and so incorrectly returned a no
658 match.
660 17. There was optimizing code for the last branch of non-capturing parentheses,
661 and also for the obeyed branch of a conditional subexpression, which used
662 tail recursion to cut down on stack usage. Unfortunately, now that there is
663 the possibility of (*THEN) occurring in these branches, tail recursion is
664 no longer possible because the return has to be checked for (*THEN). These
665 two optimizations have therefore been removed. [But see 8.20/11 above.]
667 18. If a pattern containing \R was studied, it was assumed that \R always
668 matched two bytes, thus causing the minimum subject length to be
669 incorrectly computed because \R can also match just one byte.
671 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
672 was incorrectly computed.
674 20. If /S is present twice on a test pattern in pcretest input, it now
675 *disables* studying, thereby overriding the use of -s on the command line
676 (see 14 above). This is necessary for one or two tests to keep the output
677 identical in both cases.
679 21. When (*ACCEPT) was used in an assertion that matched an empty string and
680 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
682 22. When an atomic group that contained a capturing parenthesis was
683 successfully matched, but the branch in which it appeared failed, the
684 capturing was not being forgotten if a higher numbered group was later
685 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
686 group 1 to "a", when in fact it should be unset. This applied to multi-
687 branched capturing and non-capturing groups, repeated or not, and also to
688 positive assertions (capturing in negative assertions does not happen
689 in PCRE) and also to nested atomic groups.
691 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
692 subject after a captured substring, to make it easier to tell which of a
693 number of identical substrings has been captured.
695 24. The way atomic groups are processed by pcre_exec() has been changed so that
696 if they are repeated, backtracking one repetition now resets captured
697 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
698 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
699 "aaa". Previously, it would have been "a". As part of this code
700 refactoring, the way recursive calls are handled has also been changed.
702 25. If an assertion condition captured any substrings, they were not passed
703 back unless some other capturing happened later. For example, if
704 (?(?=(a))a) was matched against "a", no capturing was returned.
706 26. When studying a pattern that contained subroutine calls or assertions,
707 the code for finding the minimum length of a possible match was handling
708 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
709 group 1 called group 2 while simultaneously a separate group 2 called group
710 1). A stack overflow occurred in this case. I have fixed this by limiting
711 the recursion depth to 10.
713 27. Updated RunTest.bat in the distribution to the version supplied by Tom
714 Fortmann. This supports explicit test numbers on the command line, and has
715 argument validation and error reporting.
717 28. An instance of \X with an unlimited repeat could fail if at any point the
718 first character it looked at was a mark character.
720 29. Some minor code refactoring concerning Unicode properties and scripts
721 should reduce the stack requirement of match() slightly.
723 30. Added the '=' option to pcretest to check the setting of unused capturing
724 slots at the end of the pattern, which are documented as being -1, but are
725 not included in the return count.
727 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
728 compiled something random. Now it gives a compile-time error (as does
729 Perl).
731 32. A *MARK encountered during the processing of a positive assertion is now
732 recorded and passed back (compatible with Perl).
734 33. If --only-matching or --colour was set on a pcregrep call whose pattern
735 had alternative anchored branches, the search for a second match in a line
736 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
737 matched the line "0102" twice. The same bug affected patterns that started
738 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
739 twice.
741 34. Previously, PCRE did not allow quantification of assertions. However, Perl
742 does, and because of capturing effects, quantifying parenthesized
743 assertions may at times be useful. Quantifiers are now allowed for
744 parenthesized assertions.
746 35. A minor code tidy in pcre_compile() when checking options for \R usage.
748 36. \g was being checked for fancy things in a character class, when it should
749 just be a literal "g".
751 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
752 appearance of a nested POSIX class supersedes an apparent external class.
753 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
754 unescaped square brackets may also appear as part of class names. For
755 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
756 more like Perl. (But see 8.20/1 above.)
758 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
759 was because it thought it was \N{name}, which is not supported).
761 39. Add minix to OS list not supporting the -S option in pcretest.
763 40. PCRE tries to detect cases of infinite recursion at compile time, but it
764 cannot analyze patterns in sufficient detail to catch mutual recursions
765 such as ((?1))((?2)). There is now a runtime test that gives an error if a
766 subgroup is called recursively as a subpattern for a second time at the
767 same position in the subject string. In previous releases this might have
768 been caught by the recursion limit, or it might have run out of stack.
770 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
771 happen only once. PCRE was, however incorrectly giving a compile time error
772 "recursive call could loop indefinitely" because it cannot analyze the
773 pattern in sufficient detail. The compile time test no longer happens when
774 PCRE is compiling a conditional subpattern, but actual runaway loops are
775 now caught at runtime (see 40 above).
777 42. It seems that Perl allows any characters other than a closing parenthesis
778 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
779 has been changed to be the same.
781 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
782 as not to get warnings when autogen.sh is called. Also changed
783 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
785 44. To help people who use pcregrep to scan files containing exceedingly long
786 lines, the following changes have been made:
788 (a) The default value of the buffer size parameter has been increased from
789 8K to 20K. (The actual buffer used is three times this size.)
791 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
792 PCRE is built.
794 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
795 to be set at run time.
797 (d) Numerical values in pcregrep options can be followed by K or M, for
798 example --buffer-size=50K.
800 (e) If a line being scanned overflows pcregrep's buffer, an error is now
801 given and the return code is set to 2.
803 45. Add a pointer to the latest mark to the callout data block.
805 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
806 partial match of an empty string instead of no match. This was specific to
807 the use of ".".
809 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
810 complete match instead of a partial match. This bug was dependent on both
811 the PCRE_UTF8 and PCRE_DOTALL options being set.
813 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
814 starting byte set, because \b was not being ignored.
817 Version 8.12 15-Jan-2011
818 ------------------------
820 1. Fixed some typos in the markup of the man pages, and wrote a script that
821 checks for such things as part of the documentation building process.
823 2. On a big-endian 64-bit system, pcregrep did not correctly process the
824 --match-limit and --recursion-limit options (added for 8.11). In
825 particular, this made one of the standard tests fail. (The integer value
826 went into the wrong half of a long int.)
828 3. If the --colour option was given to pcregrep with -v (invert match), it
829 did strange things, either producing crazy output, or crashing. It should,
830 of course, ignore a request for colour when reporting lines that do not
831 match.
833 4. Another pcregrep bug caused similar problems if --colour was specified with
834 -M (multiline) and the pattern match finished with a line ending.
836 5. In pcregrep, when a pattern that ended with a literal newline sequence was
837 matched in multiline mode, the following line was shown as part of the
838 match. This seems wrong, so I have changed it.
840 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
841 the check for further matches in the same line (so they could be coloured)
842 to overrun the end of the current line. If another match was found, it was
843 incorrectly shown (and then shown again when found in the next line).
845 7. If pcregrep was compiled under Windows, there was a reference to the
846 function pcregrep_exit() before it was defined. I am assuming this was
847 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
848 reported by a user. I've moved the definition above the reference.
851 Version 8.11 10-Dec-2010
852 ------------------------
854 1. (*THEN) was not working properly if there were untried alternatives prior
855 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
856 backtracked to try for "b" instead of moving to the next alternative branch
857 at the same level (in this case, to look for "c"). The Perl documentation
858 is clear that when (*THEN) is backtracked onto, it goes to the "next
859 alternative in the innermost enclosing group".
861 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
862 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
863 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
864 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
865 (*THEN).
867 3. If \s appeared in a character class, it removed the VT character from
868 the class, even if it had been included by some previous item, for example
869 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
870 of \s, but is part of the POSIX "space" class.)
872 4. A partial match never returns an empty string (because you can always
873 match an empty string at the end of the subject); however the checking for
874 an empty string was starting at the "start of match" point. This has been
875 changed to the "earliest inspected character" point, because the returned
876 data for a partial match starts at this character. This means that, for
877 example, /(?<=abc)def/ gives a partial match for the subject "abc"
878 (previously it gave "no match").
880 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
881 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
882 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
883 has an implication that the given string is incomplete (because a partial
884 match is preferred over a full match). For this reason, these items now
885 give a partial match in this situation. [Aside: previously, the one case
886 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
887 match rather than a full match, which was wrong by the old rules, but is
888 now correct.]
890 6. There was a bug in the handling of #-introduced comments, recognized when
891 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
892 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
893 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
894 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
895 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
896 places in pcre_compile().
898 7. Related to (6) above, when pcre_compile() was skipping #-introduced
899 comments when looking ahead for named forward references to subpatterns,
900 the only newline sequence it recognized was NL. It now handles newlines
901 according to the set newline convention.
903 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
904 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
905 cater for a lack of strerror(). These oversights have been fixed.
907 9. Added --match-limit and --recursion-limit to pcregrep.
909 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
911 11. When the -o option was used, pcregrep was setting a return code of 1, even
912 when matches were found, and --line-buffered was not being honoured.
914 12. Added an optional parentheses number to the -o and --only-matching options
915 of pcregrep.
917 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
918 can match an empty string. The code to do it in pcretest and pcredemo
919 needed fixing:
921 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
922 only one byte after an empty string match just before CRLF (this case
923 just got forgotten; "any" and "anycrlf" were OK).
925 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
926 mode when an empty string match preceded an ASCII character followed by
927 a non-ASCII character. (The code for advancing by one character rather
928 than one byte was nonsense.)
930 (c) The pcredemo.c sample program did not have any code at all to handle
931 the cases when CRLF is a valid newline sequence.
933 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
934 as a starting offset was within the subject string. There is now a new
935 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
936 negative or greater than the length of the string. In order to test this,
937 pcretest is extended to allow the setting of negative starting offsets.
939 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
940 starting offset points to the beginning of a UTF-8 character was
941 unnecessarily clumsy. I tidied it up.
943 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
944 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
946 17. Nobody had reported that the --include_dir option, which was added in
947 release 7.7 should have been called --include-dir (hyphen, not underscore)
948 for compatibility with GNU grep. I have changed it to --include-dir, but
949 left --include_dir as an undocumented synonym, and the same for
950 --exclude-dir, though that is not available in GNU grep, at least as of
951 release 2.5.4.
953 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
954 characters from a string of bytes) have been redefined so as not to use
955 loops, in order to improve performance in some environments. At the same
956 time, I abstracted some of the common code into auxiliary macros to save
957 repetition (this should not affect the compiled code).
959 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
960 compile-time error is now given if \c is not followed by an ASCII
961 character, that is, a byte less than 128. (In EBCDIC mode, the code is
962 different, and any byte value is allowed.)
964 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
965 START_OPTIMIZE option, which is now allowed at compile time - but just
966 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
967 to pcregrep and other applications that have no direct access to PCRE
968 options. The new /Y option in pcretest sets this option when calling
969 pcre_compile().
971 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
972 back references. Groups containing recursive back references were forced to
973 be atomic by that change, but in the case of named groups, the amount of
974 memory required was incorrectly computed, leading to "Failed: internal
975 error: code overflow". This has been fixed.
977 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
978 pcretest.c, to avoid build problems in some Borland environments.
981 Version 8.10 25-Jun-2010
982 ------------------------
984 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
985 THEN.
987 2. (*ACCEPT) was not working when inside an atomic group.
989 3. Inside a character class, \B is treated as a literal by default, but
990 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
991 causes the error). The code is unchanged, but I tidied the documentation.
993 4. Inside a character class, PCRE always treated \R and \X as literals,
994 whereas Perl faults them if its -w option is set. I have changed PCRE so
995 that it faults them when PCRE_EXTRA is set.
997 5. Added support for \N, which always matches any character other than
998 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1000 6. When compiling pcregrep with newer versions of gcc which may have
1001 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1002 declared with attribute warn_unused_result" were given. Just casting the
1003 result to (void) does not stop the warnings; a more elaborate fudge is
1004 needed. I've used a macro to implement this.
1006 7. Minor change to pcretest.c to avoid a compiler warning.
1008 8. Added four artifical Unicode properties to help with an option to make
1009 \s etc use properties (see next item). The new properties are: Xan
1010 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1012 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1013 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1014 this option. Modified pcretest to add /W to test this facility. Added
1015 REG_UCP to make it available via the POSIX interface.
1017 10. Added --line-buffered to pcregrep.
1019 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1020 studied, and the match started with a letter with a code point greater than
1021 127 whose first byte was different to the first byte of the other case of
1022 the letter, the other case of this starting letter was not recognized
1023 (#976).
1025 12. If a pattern that was studied started with a repeated Unicode property
1026 test, for example, \p{Nd}+, there was the theoretical possibility of
1027 setting up an incorrect bitmap of starting bytes, but fortunately it could
1028 not have actually happened in practice until change 8 above was made (it
1029 added property types that matched character-matching opcodes).
1031 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1032 possible starting bytes for non-anchored patterns.
1034 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1035 \R, and also a number of cases that involve Unicode properties, both
1036 explicit and implicit when PCRE_UCP is set.
1038 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1039 input, it could crash or give wrong results if characters with values
1040 greater than 0xc0 were present in the subject string. (Detail: it assumed
1041 UTF-8 input when processing these items.)
1043 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1044 size_t is 64-bit (#991).
1046 17. Added a check for running out of memory when PCRE is compiled with
1047 --disable-stack-for-recursion (#990).
1049 18. If the last data line in a file for pcretest does not have a newline on
1050 the end, a newline was missing in the output.
1052 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1053 less than 128) in its various bitmaps. However, there is a facility for
1054 generating tables according to the current locale when PCRE is compiled. It
1055 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1056 characters, are recognized by isspace() and therefore were getting set in
1057 these tables, and indeed these tables seem to approximate to ISO 8859. This
1058 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1059 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1060 which of course cannot start UTF-8 characters. I have changed the code so
1061 that only real ASCII characters (less than 128) and the correct starting
1062 bytes for UTF-8 encodings are set for characters greater than 127 when in
1063 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1064 altogether.)
1066 20. Added the /T option to pcretest so as to be able to run tests with non-
1067 standard character tables, thus making it possible to include the tests
1068 used for 19 above in the standard set of tests.
1070 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1071 reference to a subpattern the other side of a comment that contains an
1072 opening parenthesis caused either an internal compiling error, or a
1073 reference to the wrong subpattern.
1076 Version 8.02 19-Mar-2010
1077 ------------------------
1079 1. The Unicode data tables have been updated to Unicode 5.2.0.
1081 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1082 configured.
1084 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1085 original author of that file, following a query about its status.
1087 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1088 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1090 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1091 quantifier applied to a forward-referencing subroutine call, could compile
1092 incorrect code or give the error "internal error: previously-checked
1093 referenced subpattern not found".
1095 6. Both MS Visual Studio and Symbian OS have problems with initializing
1096 variables to point to external functions. For these systems, therefore,
1097 pcre_malloc etc. are now initialized to local functions that call the
1098 relevant global functions.
1100 7. There were two entries missing in the vectors called coptable and poptable
1101 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1102 I've fixed the data, and added a kludgy way of testing at compile time that
1103 the lengths are correct (equal to the number of opcodes).
1105 8. Following on from 7, I added a similar kludge to check the length of the
1106 eint vector in pcreposix.c.
1108 9. Error texts for pcre_compile() are held as one long string to avoid too
1109 much relocation at load time. To find a text, the string is searched,
1110 counting zeros. There was no check for running off the end of the string,
1111 which could happen if a new error number was added without updating the
1112 string.
1114 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1116 11. \K was not working if it appeared in an atomic group or in a group that
1117 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1118 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1119 the assertion is positive, but not if it is negative.
1121 12. Change 11 fortuitously reduced the size of the stack frame used in the
1122 "match()" function of pcre_exec.c by one pointer. Forthcoming
1123 implementation of support for (*MARK) will need an extra pointer on the
1124 stack; I have reserved it now, so that the stack frame size does not
1125 decrease.
1127 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1128 item in branch that calls a recursion is a subroutine call - as in the
1129 second branch in the above example - was incorrectly given the compile-
1130 time error "recursive call could loop indefinitely" because pcre_compile()
1131 was not correctly checking the subroutine for matching a non-empty string.
1133 14. The checks for overrunning compiling workspace could trigger after an
1134 overrun had occurred. This is a "should never occur" error, but it can be
1135 triggered by pathological patterns such as hundreds of nested parentheses.
1136 The checks now trigger 100 bytes before the end of the workspace.
1138 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1141 Version 8.01 19-Jan-2010
1142 ------------------------
1144 1. If a pattern contained a conditional subpattern with only one branch (in
1145 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1146 computed the wrong minimum data length (which is of course zero for such
1147 subpatterns). This could cause incorrect "no match" results.
1149 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1150 the pattern is reset in the first branch, pcre_compile() failed with
1151 "internal error: code overflow at offset...". This happened only when
1152 the reset was to the original external option setting. (An optimization
1153 abstracts leading options settings into an external setting, which was the
1154 cause of this.)
1156 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1157 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1158 assertion pattern did not match (meaning that the assertion was true), it
1159 was incorrectly treated as false if the SKIP had been reached during the
1160 matching. This also applied to assertions used as conditions.
1162 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1163 assertion subpattern, including such a pattern used as a condition,
1164 unpredictable results occurred, instead of the error return
1167 5. The C++ GlobalReplace function was not working like Perl for the special
1168 situation when an empty string is matched. It now does the fancy magic
1169 stuff that is necessary.
1171 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1172 removed. (These were left over from very, very early versions of PCRE.)
1174 7. Some cosmetic changes to the code to make life easier when compiling it
1175 as part of something else:
1177 (a) Change DEBUG to PCRE_DEBUG.
1179 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1180 called "current" as "current_branch", to prevent a collision with the
1181 Linux macro when compiled as a kernel module.
1183 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1184 prevent a collision with the Linux macro when compiled as a kernel
1185 module.
1187 8. In pcre_compile() there are some checks for integer overflows that used to
1188 cast potentially large values to (double). This has been changed to that
1189 when building, a check for int64_t is made, and if it is found, it is used
1190 instead, thus avoiding the use of floating point arithmetic. (There is no
1191 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1192 double.
1194 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1195 2005 (difference between two addresses compared to an unsigned value).
1197 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1198 custom one, because of the following reported problem in Windows:
1200 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1201 under Win32.
1202 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1203 therefore missing the function definition.
1204 - The compiler thus generates a "C" signature for the test function.
1205 - The linker fails to find the "C" function.
1206 - PCRE fails to configure if asked to do so against libbz2.
1208 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1209 messages were output:
1211 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1212 rerunning libtoolize, to keep the correct libtool macros in-tree.
1213 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1215 I have done both of these things.
1217 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1218 most of the time, it *can* run out if it is given a pattern that contains a
1219 runaway infinite recursion. I updated the discussion in the pcrestack man
1220 page.
1222 13. Now that we have gone to the x.xx style of version numbers, the minor
1223 version may start with zero. Using 08 or 09 is a bad idea because users
1224 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1225 interpreted as invalid octal numbers. I've updated the previous comment in
1226 configure.ac, and also added a check that gives an error if 08 or 09 are
1227 used.
1229 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1230 causing partial matching to fail when the end of the subject matched \W
1231 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1233 15. There were some discrepancies between the declarations in pcre_internal.h
1234 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1235 their definitions. The declarations used "const uschar *" and the
1236 definitions used USPTR. Even though USPTR is normally defined as "const
1237 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1238 reported that: "This difference in casting confuses some C++ compilers, for
1239 example, SunCC recognizes above declarations as different functions and
1240 generates broken code for hbpcre." I have changed the declarations to use
1241 USPTR.
1243 16. GNU libtool is named differently on some systems. The autogen.sh script now
1244 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1245 (FreeBSD).
1247 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1248 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1249 comment: "Figure out how to create a longlong from a string: strtoll and
1250 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1251 instance, but it only takes 2 args instead of 3!"
1253 18. A subtle bug concerned with back references has been fixed by a change of
1254 specification, with a corresponding code fix. A pattern such as
1255 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1256 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1257 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1258 same bug. Such groups have to be quantified to be useful, or contained
1259 inside another quantified group. (If there's no repetition, the reference
1260 can never match.) The problem arises because, having left the group and
1261 moved on to the rest of the pattern, a later failure that backtracks into
1262 the group uses the captured value from the final iteration of the group
1263 rather than the correct earlier one. I have fixed this in PCRE by forcing
1264 any group that contains a reference to itself to be an atomic group; that
1265 is, there cannot be any backtracking into it once it has completed. This is
1266 similar to recursive and subroutine calls.
1269 Version 8.00 19-Oct-09
1270 ----------------------
1272 1. The table for translating pcre_compile() error codes into POSIX error codes
1273 was out-of-date, and there was no check on the pcre_compile() error code
1274 being within the table. This could lead to an OK return being given in
1275 error.
1277 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1278 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1279 in a Windows environment.
1281 3. The pcregrep --count option prints the count for each file even when it is
1282 zero, as does GNU grep. However, pcregrep was also printing all files when
1283 --files-with-matches was added. Now, when both options are given, it prints
1284 counts only for those files that have at least one match. (GNU grep just
1285 prints the file name in this circumstance, but including the count seems
1286 more useful - otherwise, why use --count?) Also ensured that the
1287 combination -clh just lists non-zero counts, with no names.
1289 4. The long form of the pcregrep -F option was incorrectly implemented as
1290 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1291 but it seems right to fix it, and I didn't think it was worth preserving
1292 the old behaviour.
1294 5. The command line items --regex=pattern and --regexp=pattern were not
1295 recognized by pcregrep, which required --regex pattern or --regexp pattern
1296 (with a space rather than an '='). The man page documented the '=' forms,
1297 which are compatible with GNU grep; these now work.
1299 6. No libpcreposix.pc file was created for pkg-config; there was just
1300 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1302 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1303 when UCP support is not needed, by modifying the Python script that
1304 generates it from Unicode data files. This should not matter if the module
1305 is correctly used as a library, but I received one complaint about 50K of
1306 unwanted data. My guess is that the person linked everything into his
1307 program rather than using a library. Anyway, it does no harm.
1309 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1310 was a minimum greater than 1 for a wide character in a possessive
1311 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1312 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1313 character. Chaos in the form of incorrect output or a compiling loop could
1314 result.
1316 9. The restrictions on what a pattern can contain when partial matching is
1317 requested for pcre_exec() have been removed. All patterns can now be
1318 partially matched by this function. In addition, if there are at least two
1319 slots in the offset vector, the offset of the earliest inspected character
1320 for the match and the offset of the end of the subject are set in them when
1321 PCRE_ERROR_PARTIAL is returned.
1323 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1324 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1325 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1326 and may be more useful for multi-segment matching.
1328 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1329 used to be given if ever the end of the subject was reached; now it is
1330 given only if matching could not proceed because another character was
1331 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1332 string "Z", which now yields "no match" instead of "partial match". In the
1333 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1334 final character ended with (*FAIL).
1336 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1337 if the pattern had a "must contain" character that was already found in the
1338 earlier partial match, unless partial matching was again requested. For
1339 example, with the pattern /dog.(body)?/, the "must contain" character is
1340 "g". If the first part-match was for the string "dog", restarting with
1341 "sbody" failed. This bug has been fixed.
1343 13. The string returned by pcre_dfa_exec() after a partial match has been
1344 changed so that it starts at the first inspected character rather than the
1345 first character of the match. This makes a difference only if the pattern
1346 starts with a lookbehind assertion or \b or \B (\K is not supported by
1347 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1348 matching functions compatible, and I think it's the right thing to do.
1350 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1351 so that the demonstration program is easily available in environments where
1352 PCRE has not been installed from source.
1354 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1355 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1356 library.
1358 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1359 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1360 is not the first non-POSIX option to be added. Clearly some people find
1361 these options useful.
1363 17. If a caller to the POSIX matching function regexec() passes a non-zero
1364 value for nmatch with a NULL value for pmatch, the value of
1365 nmatch is forced to zero.
1367 18. RunGrepTest did not have a test for the availability of the -u option of
1368 the diff command, as RunTest does. It now checks in the same way as
1369 RunTest, and also checks for the -b option.
1371 19. If an odd number of negated classes containing just a single character
1372 interposed, within parentheses, between a forward reference to a named
1373 subpattern and the definition of the subpattern, compilation crashed with
1374 an internal error, complaining that it could not find the referenced
1375 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1376 [The bug was that it was starting one character too far in when skipping
1377 over the character class, thus treating the ] as data rather than
1378 terminating the class. This meant it could skip too much.]
1380 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1381 /g option in pcretest when the pattern contains \K, which makes it possible
1382 to have an empty string match not at the start, even when the pattern is
1383 anchored. Updated pcretest and pcredemo to use this option.
1385 21. If the maximum number of capturing subpatterns in a recursion was greater
1386 than the maximum at the outer level, the higher number was returned, but
1387 with unset values at the outer level. The correct (outer level) value is
1388 now given.
1390 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1391 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1392 make it do so. The string so far is captured, making this feature
1393 compatible with Perl.
1395 23. The tests have been re-organized, adding tests 11 and 12, to make it
1396 possible to check the Perl 5.10 features against Perl 5.10.
1398 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1399 pattern matches a fixed length string. PCRE did not allow this; now it
1400 does. Neither allows recursion.
1402 25. I finally figured out how to implement a request to provide the minimum
1403 length of subject string that was needed in order to match a given pattern.
1404 (It was back references and recursion that I had previously got hung up
1405 on.) This code has now been added to pcre_study(); it finds a lower bound
1406 to the length of subject needed. It is not necessarily the greatest lower
1407 bound, but using it to avoid searching strings that are too short does give
1408 some useful speed-ups. The value is available to calling programs via
1409 pcre_fullinfo().
1411 26. While implementing 25, I discovered to my embarrassment that pcretest had
1412 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1413 study optimizations had never been tested with that matching function.
1414 Oops. What is worse, even when it was passed study data, there was a bug in
1415 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1416 were also very few tests of studied patterns with pcre_dfa_exec().
1418 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1419 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1420 on the other side of the coin, they are no longer allowed to have different
1421 names, because these cannot be distinguished in PCRE, and this has caused
1422 confusion. (This is a difference from Perl.)
1424 28. When duplicate subpattern names are present (necessarily with different
1425 numbers, as required by 27 above), and a test is made by name in a
1426 conditional pattern, either for a subpattern having been matched, or for
1427 recursion in such a pattern, all the associated numbered subpatterns are
1428 tested, and the overall condition is true if the condition is true for any
1429 one of them. This is the way Perl works, and is also more like the way
1430 testing by number works.
1433 Version 7.9 11-Apr-09
1434 ---------------------
1436 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1437 (pcretest), all targets were linked against these libraries. This included
1438 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1439 libraries. This caused unwanted dependencies to be created. This problem
1440 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1441 pcretest is linked with readline.
1443 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1444 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1445 moved outside it again, because FALSE and TRUE are already defined in AIX,
1446 but BOOL is not.
1448 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1449 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1451 4. The pcregrep documentation said spaces were inserted as well as colons (or
1452 hyphens) following file names and line numbers when outputting matching
1453 lines. This is not true; no spaces are inserted. I have also clarified the
1454 wording for the --colour (or --color) option.
1456 5. In pcregrep, when --colour was used with -o, the list of matching strings
1457 was not coloured; this is different to GNU grep, so I have changed it to be
1458 the same.
1460 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1461 each matching line was coloured. Now it goes on to look for further matches
1462 of any of the test patterns, which is the same behaviour as GNU grep.
1464 7. A pattern that could match an empty string could cause pcregrep to loop; it
1465 doesn't make sense to accept an empty string match in pcregrep, so I have
1466 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1467 seems to be how GNU grep behaves.
1469 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1470 start or after a newline", because the conditional assertion was not being
1471 correctly handled. The rule now is that both the assertion and what follows
1472 in the first alternative must satisfy the test.
1474 9. If auto-callout was enabled in a pattern with a conditional group whose
1475 condition was an assertion, PCRE could crash during matching, both with
1476 pcre_exec() and pcre_dfa_exec().
1478 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1479 used for matching.
1481 11. Unicode property support in character classes was not working for
1482 characters (bytes) greater than 127 when not in UTF-8 mode.
1484 12. Added the -M command line option to pcretest.
1486 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1488 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1490 16. Added comments and documentation about mis-use of no_arg in the C++
1491 wrapper.
1493 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1494 from Martin Jerabek that uses macro names for all relevant character and
1495 string constants.
1497 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1498 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1499 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1500 these, but not everybody uses configure.
1502 19. A conditional group that had only one branch was not being correctly
1503 recognized as an item that could match an empty string. This meant that an
1504 enclosing group might also not be so recognized, causing infinite looping
1505 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1506 with the subject "ab", where knowledge that the repeated group can match
1507 nothing is needed in order to break the loop.
1509 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1510 exec(), but without supplying a callout function, matching went wrong.
1512 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1513 leak if the size of the offset vector was greater than 30. When the vector
1514 is smaller, the saved offsets during recursion go onto a local stack
1515 vector, but for larger vectors malloc() is used. It was failing to free
1516 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1517 error, in fact).
1519 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1520 heapframe that is used only when UTF-8 support is enabled. This caused no
1521 problem, but was untidy.
1523 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1524 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1525 included within another project.
1527 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1528 slightly modified by me:
1530 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1531 not building pcregrep.
1533 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1534 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1536 25. Forward references, both numeric and by name, in patterns that made use of
1537 duplicate group numbers, could behave incorrectly or give incorrect errors,
1538 because when scanning forward to find the reference group, PCRE was not
1539 taking into account the duplicate group numbers. A pattern such as
1540 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1542 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1543 the feature of a custom pointer more persuasive (as requested by a user).
1545 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1546 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1547 pre-defined.
1549 28. Added support for (*UTF8) at the start of a pattern.
1551 29. Arrange for flags added by the "release type" setting in CMake to be shown
1552 in the configuration summary.
1555 Version 7.8 05-Sep-08
1556 ---------------------
1558 1. Replaced UCP searching code with optimized version as implemented for Ad
1559 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1560 stage table and inline lookup instead of a function, giving speed ups of 2
1561 to 5 times on some simple patterns that I tested. Permission was given to
1562 distribute the MultiStage2.py script that generates the tables (it's not in
1563 the tarball, but is in the Subversion repository).
1565 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1566 scripts.
1568 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1569 a group with a zero qualifier. The result of the study could be incorrect,
1570 or the function might crash, depending on the pattern.
1572 4. Caseless matching was not working for non-ASCII characters in back
1573 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1574 It now works when Unicode Property Support is available.
1576 5. In pcretest, an escape such as \x{de} in the data was always generating
1577 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1578 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1579 truncation.
1581 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1583 7. Added two (int) casts to pcregrep when printing the difference of two
1584 pointers, in case they are 64-bit values.
1586 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1587 test 2 if it fails.
1589 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1590 and a #define of that name to empty if it is not externally set. This is to
1591 allow users of MSVC to set it if necessary.
1593 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1594 the convenience functions in the pcre_get.c source file.
1596 11. An option change at the start of a pattern that had top-level alternatives
1597 could cause overwriting and/or a crash. This command provoked a crash in
1598 some environments:
1600 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1602 This potential security problem was recorded as CVE-2008-2371.
1604 12. For a pattern where the match had to start at the beginning or immediately
1605 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1606 pcre_dfa_exec() could read past the end of the passed subject if there was
1607 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1608 pcretest so that it places the subject at the end of its malloc-ed buffer.
1610 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1611 exec() might read past the end of the data buffer in UTF-8 mode.
1613 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1614 the data contained the byte 0x85 as part of a UTF-8 character within its
1615 first line. This applied both to normal and DFA matching.
1617 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1618 /^[^d]*?$/8 failed to match "abc".
1620 16. Added a missing copyright notice to pcrecpp_internal.h.
1622 17. Make it more clear in the documentation that values returned from
1623 pcre_exec() in ovector are byte offsets, not character counts.
1625 18. Tidied a few places to stop certain compilers from issuing warnings.
1627 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1628 supplied by Stefan Weber. I made a further small update for 7.8 because
1629 there is a change of source arrangements: the pcre_searchfuncs.c module is
1630 replaced by pcre_ucd.c.
1633 Version 7.7 07-May-08
1634 ---------------------
1636 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1637 a string to a long long, pretend we don't even have a long long." This is
1638 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1640 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1641 pre-7.6 versions, which defined a global no_arg variable instead of putting
1642 it in the RE class. (See also #8 below.)
1644 3. Remove a line of dead code, identified by coverity and reported by Nuno
1645 Lopes.
1647 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1649 (1) The include/exclude patterns were being applied to the whole pathnames
1650 of files, instead of just to the final components.
1652 (2) If there was more than one level of directory, the subdirectories were
1653 skipped unless they satisfied the include/exclude conditions. This is
1654 inconsistent with GNU grep (and could even be seen as contrary to the
1655 pcregrep specification - which I improved to make it absolutely clear).
1656 The action now is always to scan all levels of directory, and just
1657 apply the include/exclude patterns to regular files.
1659 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1660 --exclude_dir in the tests to avoid scanning .svn directories.
1662 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1663 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1664 doesn't support NULs in patterns.
1666 7. Added some missing "const"s to declarations of static tables in
1667 pcre_compile.c and pcre_dfa_exec.c.
1669 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1670 caused by fix #2 above. (Subsequently also a second patch to fix the
1671 first patch. And a third patch - this was a messy problem.)
1673 9. Applied Craig's patch to remove the use of push_back().
1675 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1676 matching function regexec().
1678 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1679 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1680 references. PCRE supports relative numbers with this syntax (I don't think
1681 Oniguruma does).
1683 12. Previously, a group with a zero repeat such as (...){0} was completely
1684 omitted from the compiled regex. However, this means that if the group
1685 was called as a subroutine from elsewhere in the pattern, things went wrong
1686 (an internal error was given). Such groups are now left in the compiled
1687 pattern, with a new opcode that causes them to be skipped at execution
1688 time.
1690 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1691 to the way PCRE behaves:
1693 (a) A lone ] character is dis-allowed (Perl treats it as data).
1695 (b) A back reference to an unmatched subpattern matches an empty string
1696 (Perl fails the current match path).
1698 (c) A data ] in a character class must be notated as \] because if the
1699 first data character in a class is ], it defines an empty class. (In
1700 Perl it is not possible to have an empty class.) The empty class []
1701 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1702 The negative empty class [^] matches any one character, independently
1703 of the DOTALL setting.
1705 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1706 non-existent subpattern following a character class starting with ']' and
1707 containing () gave an internal compiling error instead of "reference to
1708 non-existent subpattern". Fortunately, when the pattern did exist, the
1709 compiled code was correct. (When scanning forwards to check for the
1710 existencd of the subpattern, it was treating the data ']' as terminating
1711 the class, so got the count wrong. When actually compiling, the reference
1712 was subsequently set up correctly.)
1714 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1715 it was being rejected as not supported by pcre_dfa_exec(), even though
1716 other assertions are supported. I have made pcre_dfa_exec() support
1717 (*FAIL).
1719 16. The implementation of 13c above involved the invention of a new opcode,
1720 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1721 cannot be changed at match time, I realized I could make a small
1722 improvement to matching performance by compiling OP_ALLANY instead of
1723 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1724 on the OP_ANY path.
1726 17. Compiling pcretest on Windows with readline support failed without the
1727 following two fixes: (1) Make the unistd.h include conditional on
1728 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1730 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1731 ncurses library to be included for pcretest when ReadLine support is
1732 requested, but also to allow for it to be overridden. This patch came from
1733 Daniel Bergström.
1735 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1736 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1737 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1738 spotting this.
1741 Version 7.6 28-Jan-08
1742 ---------------------
1744 1. A character class containing a very large number of characters with
1745 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1746 overflow.
1748 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1749 HAVE_LONG_LONG is not defined.
1751 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1752 bring them up to date and include new features. This patch includes:
1754 - Fixed PH's badly added libz and libbz2 support.
1755 - Fixed a problem with static linking.
1756 - Added pcredemo. [But later removed - see 7 below.]
1757 - Fixed dftables problem and added an option.
1758 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1760 - Added readline support for pcretest.
1761 - Added an listing of the option settings after cmake has run.
1763 4. A user submitted a patch to Makefile that makes it easy to create
1764 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1765 Makefile.am that cause it to include this special target, without
1766 affecting anything else. Note that the same mingw target plus all
1767 the other distribution libraries and programs are now supported
1768 when configuring with CMake (see 6 below) instead of with
1769 Configure/Make.
1771 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1772 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1773 exported in the Windows port". It has not yet been confirmed that the patch
1774 solves the problem, but it does no harm.
1776 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1777 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1778 with CMake, and also correct the comment about stack recursion.
1780 7. Remove the automatic building of pcredemo from the ./configure system and
1781 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1782 of a program that users should build themselves after PCRE is installed, so
1783 building it automatically is not really right. What is more, it gave
1784 trouble in some build environments.
1786 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1789 Version 7.5 10-Jan-08
1790 ---------------------
1792 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1793 values in parens when parsing an RE using the C++ wrapper."
1795 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1796 Characters greater than 255 were excluded from the class instead of being
1797 included.
1799 3. The same bug as (2) above applied to negated POSIX classes such as
1800 [:^space:].
1802 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1803 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1804 I have changed it.
1806 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1807 first named subpattern) and a construct such as (?&a) would reference the
1808 first named subpattern whose name started with "a" (in other words, the
1809 length check was missing). Both these problems are fixed. "Subpattern name
1810 expected" is now given for (?&) (a zero-length name), and this patch also
1811 makes it give the same error for \k'' (previously it complained that that
1812 was a reference to a non-existent subpattern).
1814 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1815 this is right because (?- can be followed by option settings as well as by
1816 digits. I have, however, made the messages clearer.
1818 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1819 than the number used in the conditional) now cause a compile-time error.
1820 This is actually not compatible with Perl, which accepts such patterns, but
1821 treats the conditional as always being FALSE (as PCRE used to), but it
1822 seems to me that giving a diagnostic is better.
1824 8. Change "alphameric" to the more common word "alphanumeric" in comments
1825 and messages.
1827 9. Fix two occurrences of "backslash" in comments that should have been
1828 "backspace".
1830 10. Remove two redundant lines of code that can never be obeyed (their function
1831 was moved elsewhere).
1833 11. The program that makes PCRE's Unicode character property table had a bug
1834 which caused it to generate incorrect table entries for sequences of
1835 characters that have the same character type, but are in different scripts.
1836 It amalgamated them into a single range, with the script of the first of
1837 them. In other words, some characters were in the wrong script. There were
1838 thirteen such cases, affecting characters in the following ranges:
1840 U+002b0 - U+002c1
1841 U+0060c - U+0060d
1842 U+0061e - U+00612
1843 U+0064b - U+0065e
1844 U+0074d - U+0076d
1845 U+01800 - U+01805
1846 U+01d00 - U+01d77
1847 U+01d9b - U+01dbf
1848 U+0200b - U+0200f
1849 U+030fc - U+030fe
1850 U+03260 - U+0327f
1851 U+0fb46 - U+0fbb1
1852 U+10450 - U+1049d
1854 12. The -o option (show only the matching part of a line) for pcregrep was not
1855 compatible with GNU grep in that, if there was more than one match in a
1856 line, it showed only the first of them. It now behaves in the same way as
1857 GNU grep.
1859 13. If the -o and -v options were combined for pcregrep, it printed a blank
1860 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1861 does the same. The return code can be used to tell if there were any
1862 non-matching lines.
1864 14. Added --file-offsets and --line-offsets to pcregrep.
1866 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1867 infinitely looping recursion. The bug was that positive lookaheads were not
1868 being skipped when checking for a possible empty match (negative lookaheads
1869 and both kinds of lookbehind were skipped).
1871 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1872 inclusion of <windows.h> to before rather than after the definition of
1873 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1875 17. Specifying a possessive quantifier with a specific limit for a Unicode
1876 character property caused pcre_compile() to compile bad code, which led at
1877 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1878 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1879 caused the error; without that there was no problem.
1881 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1883 19. Added --enable-pcretest-libreadline.
1885 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1886 RE::GlobalReplace(). As a result, the number of replacements returned was
1887 double what it should be. I removed one of the increments, but Craig sent a
1888 later patch that removed the other one (the right fix) and added unit tests
1889 that check the return values (which was not done before).
1891 21. Several CMake things:
1893 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1894 the names libpcre and libpcreposix, not just pcre and pcreposix.
1896 (2) The above change means that pcretest and pcregrep are now correctly
1897 linked with the newly-built libraries, not previously installed ones.
1901 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1902 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1903 UTF-8 newline character). The key issue is that the pattern starts .*;
1904 this means that the match must be either at the beginning, or after a
1905 newline. The bug was in the code for advancing after a failed match and
1906 checking that the new position followed a newline. It was not taking
1907 account of UTF-8 characters correctly.
1909 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1910 character classes. PCRE was not treating the sequence [:...:] as a
1911 character class unless the ... were all letters. Perl, however, seems to
1912 allow any characters between [: and :], though of course it rejects as
1913 unknown any "names" that contain non-letters, because all the known class
1914 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1915 for example, whereas PCRE did not - it did not recognize a POSIX character
1916 class. This seemed a bit dangerous, so the code has been changed to be
1917 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1918 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1919 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1920 Perl does, and where it didn't before.
1922 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1923 Windows environments %n is disabled by default.
1926 Version 7.4 21-Sep-07
1927 ---------------------
1929 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1930 means that a class such as [\s] counted as "explicit reference to CR or
1931 LF". That isn't really right - the whole point of the change was to try to
1932 help when there was an actual mention of one of the two characters. So now
1933 the change happens only if \r or \n (or a literal CR or LF) character is
1934 encountered.
1936 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1937 of both had grown to the point where there were only 3 bits left.
1938 Fortunately, there was spare space in the data structure, and so I have
1939 moved the internal flags into a new 16-bit field to free up more option
1940 bits.
1942 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1943 but did not set the internal JCHANGED flag - either of these is enough to
1944 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1945 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1946 start sets both bits.
1948 4. Added options (at build time, compile time, exec time) to change \R from
1949 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1951 5. doc/pcresyntax.html was missing from the distribution.
1953 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1954 compatibility, even though it is no longer used.
1956 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1957 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1958 windows.h file is present (where different names are used). [This was
1959 reversed later after testing - see 16 below.]
1961 8. Changed all #include <config.h> to #include "config.h". There were also
1962 some further <pcre.h> cases that I changed to "pcre.h".
1964 9. When pcregrep was used with the --colour option, it missed the line ending
1965 sequence off the lines that it output.
1967 10. It was pointed out to me that arrays of string pointers cause lots of
1968 relocations when a shared library is dynamically loaded. A technique of
1969 using a single long string with a table of offsets can drastically reduce
1970 these. I have refactored PCRE in four places to do this. The result is
1971 dramatic:
1973 Originally: 290
1974 After changing UCP table: 187
1975 After changing error message table: 43
1976 After changing table of "verbs" 36
1977 After changing table of Posix names 22
1979 Thanks to the folks working on Gregex for glib for this insight.
1981 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1982 unicode-properties was also set.
1984 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1986 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1987 checked only for CRLF.
1989 14. Added casts to pcretest.c to avoid compiler warnings.
1991 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1993 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1994 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1995 entirely. This removes changes made in 7 above.
1997 17. The CMake files have been updated, and there is now more information about
1998 building with CMake in the NON-UNIX-USE document.
2001 Version 7.3 28-Aug-07
2002 ---------------------
2004 1. In the rejigging of the build system that eventually resulted in 7.1, the
2005 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2006 brackets there is not right, since it causes compilers to look for an
2007 installed pcre.h, not the version that is in the source that is being
2008 compiled (which of course may be different). I have changed it back to:
2010 #include "pcre.h"
2012 I have a vague recollection that the change was concerned with compiling in
2013 different directories, but in the new build system, that is taken care of
2014 by the VPATH setting the Makefile.
2016 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2017 when the subject happened to end in the byte 0x85 (e.g. if the last
2018 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2019 characters but of course it shouldn't be taken as a newline when it is part
2020 of another character. The bug was that, for an unlimited repeat of . in
2021 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2022 characters when looking for a newline.
2024 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2026 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2027 in debug output.
2029 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2030 long printing in the pcrecpp unittest when running under MinGW.
2032 6. ESC_K was left out of the EBCDIC table.
2034 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2035 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2036 limit also applies to "virtual nesting" when a pattern is recursive, and in
2037 this case 1000 isn't so big. I have been able to remove this limit at the
2038 expense of backing off one optimization in certain circumstances. Normally,
2039 when pcre_exec() would call its internal match() function recursively and
2040 immediately return the result unconditionally, it uses a "tail recursion"
2041 feature to save stack. However, when a subpattern that can match an empty
2042 string has an unlimited repetition quantifier, it no longer makes this
2043 optimization. That gives it a stack frame in which to save the data for
2044 checking that an empty string has been matched. Previously this was taken
2045 from the 1000-entry workspace that had been reserved. So now there is no
2046 explicit limit, but more stack is used.
2048 8. Applied Daniel's patches to solve problems with the import/export magic
2049 syntax that is required for Windows, and which was going wrong for the
2050 pcreposix and pcrecpp parts of the library. These were overlooked when this
2051 problem was solved for the main library.
2053 9. There were some crude static tests to avoid integer overflow when computing
2054 the size of patterns that contain repeated groups with explicit upper
2055 limits. As the maximum quantifier is 65535, the maximum group length was
2056 set at 30,000 so that the product of these two numbers did not overflow a
2057 32-bit integer. However, it turns out that people want to use groups that
2058 are longer than 30,000 bytes (though not repeat them that many times).
2059 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2060 made it possible to implement the integer overflow checks in a much more
2061 dynamic way, which I have now done. The artificial limitation on group
2062 length has been removed - we now have only the limit on the total length of
2063 the compiled pattern, which depends on the LINK_SIZE setting.
2065 10. Fixed a bug in the documentation for get/copy named substring when
2066 duplicate names are permitted. If none of the named substrings are set, the
2067 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2068 empty string.
2070 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2071 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2072 because the ] is interpreted as the first data character and the
2073 terminating ] is not found. PCRE has been made compatible with Perl in this
2074 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2075 cause memory overwriting.
2077 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2078 string has been matched (to stop an infinite loop). It was not recognizing
2079 a conditional subpattern that could match an empty string if that
2080 subpattern was within another subpattern. For example, it looped when
2081 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2082 condition was not nested. This bug has been fixed.
2084 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2085 past the start of the subject in the presence of bytes with the top bit
2086 set, for example "\x8aBCD".
2088 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2089 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2091 14. Optimized (?!) to (*FAIL).
2093 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2094 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2095 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2096 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2097 does: it's just the validity check that is more restrictive.
2099 16. Inserted checks for integer overflows during escape sequence (backslash)
2100 processing, and also fixed erroneous offset values for syntax errors during
2101 backslash processing.
2103 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2104 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2106 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2107 caused an overrun.
2109 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2110 something other than just ASCII characters) inside a group that had an
2111 unlimited repeat caused a loop at compile time (while checking to see
2112 whether the group could match an empty string).
2114 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2115 [\P{Any}] did so. (Error in the code for printing property names.)
2117 21. An orphan \E inside a character class could cause a crash.
2119 22. A repeated capturing bracket such as (A)? could cause a wild memory
2120 reference during compilation.
2122 23. There are several functions in pcre_compile() that scan along a compiled
2123 expression for various reasons (e.g. to see if it's fixed length for look
2124 behind). There were bugs in these functions when a repeated \p or \P was
2125 present in the pattern. These operators have additional parameters compared
2126 with \d, etc, and these were not being taken into account when moving along
2127 the compiled data. Specifically:
2129 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2130 length.
2132 (b) An item such as \pL+ within a repeated group could cause crashes or
2133 loops.
2135 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2136 "reference to non-existent subpattern" error.
2138 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2140 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2141 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2143 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2145 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2146 character were causing crashes (broken optimization).
2148 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2149 \p or \P) caused a compile-time loop.
2151 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2152 break. For example, the unstudied pattern [\r\n]A does not match the string
2153 "\r\nA" because change 7.0/46 below moves the current point on by two
2154 characters after failing to match at the start. However, the pattern \nA
2155 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2156 the same is true. There doesn't seem any very clean way out of this, but
2157 what I have chosen to do makes the common cases work: PCRE now takes note
2158 of whether there can be an explicit match for \r or \n anywhere in the
2159 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2160 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2161 pattern has explicit CR or LF references.
2163 29. Added (*CR) etc for changing newline setting at start of pattern.
2166 Version 7.2 19-Jun-07
2167 ---------------------
2169 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2170 which is apparently normally available under Windows.
2172 2. Re-jig the pcregrep tests with different newline settings in an attempt
2173 to make them independent of the local environment's newline setting.
2175 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2177 4. Some of the "internals" tests were previously cut out when the link size
2178 was not 2, because the output contained actual offsets. The recent new
2179 "Z" feature of pcretest means that these can be cut out, making the tests
2180 usable with all link sizes.
2182 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2183 stack recursion. This gives a massive performance boost under BSD, but just
2184 a small improvement under Linux. However, it saves one field in the frame
2185 in all cases.
2187 6. Added more features from the forthcoming Perl 5.10:
2189 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2190 recursion call. It refers to the nth most recently opened parentheses.
2192 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2193 to be opened parentheses.
2195 (c) Conditions that refer to capturing parentheses can be specified
2196 relatively, for example, (?(-2)... or (?(+3)...
2198 (d) \K resets the start of the current match so that everything before
2199 is not part of it.
2201 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2203 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2204 reference syntax.
2206 (g) (?| introduces a group in which the numbering of parentheses in each
2207 alternative starts with the same number.
2209 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2211 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2214 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2215 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2216 for detecting groups that can match an empty string.
2218 9. A pattern with a very large number of alternatives (more than several
2219 hundred) was running out of internal workspace during the pre-compile
2220 phase, where pcre_compile() figures out how much memory will be needed. A
2221 bit of new cunning has reduced the workspace needed for groups with
2222 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2223 workspace instead of running out of the 4096 that are available.
2225 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2227 11. Applied patch from Google to remove an optimization that didn't quite work.
2228 The report of the bug said:
2230 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2231 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2232 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2234 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2235 it matched the wrong number of bytes.
2238 Version 7.1 24-Apr-07
2239 ---------------------
2241 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2242 that is more "standard", making use of automake and other Autotools. There
2243 is some re-arrangement of the files and adjustment of comments consequent
2244 on this.
2246 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2247 for recursive directory scanning broke on some systems because the files
2248 are not scanned in any specific order and on different systems the order
2249 was different. A call to "sort" has been inserted into RunGrepTest for the
2250 approprate test as a short-term fix. In the longer term there may be an
2251 alternative.
2253 3. I had an email from Eric Raymond about problems translating some of PCRE's
2254 man pages to HTML (despite the fact that I distribute HTML pages, some
2255 people do their own conversions for various reasons). The problems
2256 concerned the use of low-level troff macros .br and .in. I have therefore
2257 removed all such uses from the man pages (some were redundant, some could
2258 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2259 HTML has been updated to handle .nf/.fi and to complain if it encounters
2260 .br or .in.
2262 4. Updated comments in configure.ac that get placed in config.h.in and also
2263 arranged for config.h to be included in the distribution, with the name
2264 config.h.generic, for the benefit of those who have to compile without
2265 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2267 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2268 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2269 makevp.bat for the current PCRE, using the additional files
2270 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2272 6. A Windows user reported a minor discrepancy with test 2, which turned out
2273 to be caused by a trailing space on an input line that had got lost in his
2274 copy. The trailing space was an accident, so I've just removed it.
2276 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2277 that is needed.
2279 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2280 as "const" (a) because they are and (b) because it helps the PHP
2281 maintainers who have recently made a script to detect big data structures
2282 in the php code that should be moved to the .rodata section. I remembered
2283 to update Builducptable as well, so it won't revert if ucptable.h is ever
2284 re-created.
2286 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2287 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2288 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2289 support is not required. This saves 1.5-2K of code, which is important in
2290 some applications.
2292 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2293 so as not to refer to the tables, even though these functions will never be
2294 called when UTF-8 support is disabled. Otherwise there are problems with a
2295 shared library.
2297 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2299 (a) It was defining its arguments as char * instead of void *.
2301 (b) It was assuming that all moves were upwards in memory; this was true
2302 a long time ago when I wrote it, but is no longer the case.
2304 The emulated memove() is provided for those environments that have neither
2305 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2306 is clearly not the case, as these two bugs were recently reported.
2308 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2309 and Detrail to create the HTML documentation, the .txt form of the man
2310 pages, and it removes trailing spaces from listed files. It also creates
2311 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2312 case, it wraps all the #defines with #ifndefs. This script should be run
2313 before "make dist".
2315 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2316 with Unicode property support.
2318 (a) For a maximizing quantifier, if the two different cases of the
2319 character were of different lengths in their UTF-8 codings (there are
2320 some cases like this - I found 11), and the matching function had to
2321 back up over a mixture of the two cases, it incorrectly assumed they
2322 were both the same length.
2324 (b) When PCRE was configured to use the heap rather than the stack for
2325 recursion during matching, it was not correctly preserving the data for
2326 the other case of a UTF-8 character when checking ahead for a match
2327 while processing a minimizing repeat. If the check also involved
2328 matching a wide character, but failed, corruption could cause an
2329 erroneous result when trying to check for a repeat of the original
2330 character.
2332 13. Some tidying changes to the testing mechanism:
2334 (a) The RunTest script now detects the internal link size and whether there
2335 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2336 values substituted by "configure". (The RunGrepTest script already did
2337 this for UTF-8.) The configure.ac script no longer substitutes the
2338 relevant variables.
2340 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2341 with length and offset values. This means that the output is different
2342 for different internal link sizes. Test 2 is skipped for link sizes
2343 other than 2 because of this, bypassing the problem. Unfortunately,
2344 there was also a test in test 3 (the locale tests) that used /B and
2345 failed for link sizes other than 2. Rather than cut the whole test out,
2346 I have added a new /Z option to pcretest that replaces the length and
2347 offset values with spaces. This is now used to make test 3 independent
2348 of link size. (Test 2 will be tidied up later.)
2350 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2351 segmentation fault instead of returning the appropriate error message.
2353 15. In multiline mode when the newline sequence was set to "any", the pattern
2354 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2355 This doesn't seem right; it now treats the CRLF combination as the line
2356 ending, and so does not match in that case. It's only a pattern such as ^$
2357 that would hit this one: something like ^ABC$ would have failed after \r
2358 and then tried again after \r\n.
2360 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2361 in an attempt to make files that differ only in their line terminators
2362 compare equal. This works on Linux.
2364 17. Under certain error circumstances pcregrep might try to free random memory
2365 as it exited. This is now fixed, thanks to valgrind.
2367 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2368 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2369 was because its rules for how to advance for /g after matching an empty
2370 string at the end of a line did not allow for this case. They now check for
2371 it specially.
2373 20. pcretest is supposed to handle patterns and data of any length, by
2374 extending its buffers when necessary. It was getting this wrong when the
2375 buffer for a data line had to be extended.
2377 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2378 CRLF as a newline sequence.
2380 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2381 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2382 I have nevertheless tidied it up.
2384 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2386 24. Added a man page for pcre-config.
2389 Version 7.0 19-Dec-06
2390 ---------------------
2392 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2393 moving to gcc 4.1.1.
2395 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2396 sys/time.h, which is documented as needed for this function. It doesn't
2397 seem to matter on Linux, but it showed up on some releases of OS X.
2399 3. It seems that there are systems where bytes whose values are greater than
2400 127 match isprint() in the "C" locale. The "C" locale should be the
2401 default when a C program starts up. In most systems, only ASCII printing
2402 characters match isprint(). This difference caused the output from pcretest
2403 to vary, making some of the tests fail. I have changed pcretest so that:
2405 (a) When it is outputting text in the compiled version of a pattern, bytes
2406 other than 32-126 are always shown as hex escapes.
2408 (b) When it is outputting text that is a matched part of a subject string,
2409 it does the same, unless a different locale has been set for the match
2410 (using the /L modifier). In this case, it uses isprint() to decide.
2412 4. Fixed a major bug that caused incorrect computation of the amount of memory
2413 required for a compiled pattern when options that changed within the
2414 pattern affected the logic of the preliminary scan that determines the
2415 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2416 that the computed length was too small. The symptoms of this bug were
2417 either the PCRE error "internal error: code overflow" from pcre_compile(),
2418 or a glibc crash with a message such as "pcretest: free(): invalid next
2419 size (fast)". Examples of patterns that provoked this bug (shown in
2420 pcretest format) are:
2422 /(?-x: )/x
2423 /(?x)(?-x: \s*#\s*)/
2424 /((?i)[\x{c0}])/8
2425 /(?i:[\x{c0}])/8
2427 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2428 is now done differently.
2430 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2431 wrapper classes; (b) implement a new function in the C++ scanner that is
2432 more efficient than the old way of doing things because it avoids levels of
2433 recursion in the regex matching; (c) add a paragraph to the documentation
2434 for the FullMatch() function.
2436 6. The escape sequence \n was being treated as whatever was defined as
2437 "newline". Not only was this contrary to the documentation, which states
2438 that \n is character 10 (hex 0A), but it also went horribly wrong when
2439 "newline" was defined as CRLF. This has been fixed.
2441 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2442 was being set to -1 for the "end of line" case (supposedly a value that no
2443 character can have). Though this value is never used (the check for end of
2444 line is "zero bytes in current character"), it caused compiler complaints.
2445 I've changed it to 0xffffffff.
2447 8. In pcre_version.c, the version string was being built by a sequence of
2448 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2449 string (as it is for production releases) called a macro with an empty
2450 argument. The C standard says the result of this is undefined. The gcc
2451 compiler treats it as an empty string (which was what was wanted) but it is
2452 reported that Visual C gives an error. The source has been hacked around to
2453 avoid this problem.
2455 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2456 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2457 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2458 of them did).
2460 10. Originally, pcretest opened its input and output without "b"; then I was
2461 told that "b" was needed in some environments, so it was added for release
2462 5.0 to both the input and output. (It makes no difference on Unix-like
2463 systems.) Later I was told that it is wrong for the input on Windows. I've
2464 now abstracted the modes into two macros, to make it easier to fiddle with
2465 them, and removed "b" from the input mode under Windows.
2467 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2469 12. Added -help and --help to pcretest as an official way of being reminded
2470 of the options.
2472 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2473 and pcrecpp.cc because they annoy compilers at high warning levels.
2475 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2477 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2478 scripts are not C programs :-) and which was not noticed because it works
2479 on Linux.
2481 16. pcretest is supposed to handle any length of pattern and data line (as one
2482 line or as a continued sequence of lines) by extending its input buffer if
2483 necessary. This feature was broken for very long pattern lines, leading to
2484 a string of junk being passed to pcre_compile() if the pattern was longer
2485 than about 50K.
2487 17. I have done a major re-factoring of the way pcre_compile() computes the
2488 amount of memory needed for a compiled pattern. Previously, there was code
2489 that made a preliminary scan of the pattern in order to do this. That was
2490 OK when PCRE was new, but as the facilities have expanded, it has become
2491 harder and harder to keep it in step with the real compile phase, and there
2492 have been a number of bugs (see for example, 4 above). I have now found a
2493 cunning way of running the real compile function in a "fake" mode that
2494 enables it to compute how much memory it would need, while actually only
2495 ever using a few hundred bytes of working memory and without too many
2496 tests of the mode. This should make future maintenance and development
2497 easier. A side effect of this work is that the limit of 200 on the nesting
2498 depth of parentheses has been removed (though this was never a serious
2499 limitation, I suspect). However, there is a downside: pcre_compile() now
2500 runs more slowly than before (30% or more, depending on the pattern). I
2501 hope this isn't a big issue. There is no effect on runtime performance.
2503 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2504 newline (only possible for the last line of a file) and it was a
2505 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2507 19. Added additional timing features to pcretest. (1) The -tm option now times
2508 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2509 separate command line item, by a number that specifies the number of
2510 repeats to use when timing. The default is 50000; this gives better
2511 precision, but takes uncomfortably long for very large patterns.
2513 20. Extended pcre_study() to be more clever in cases where a branch of a
2514 subpattern has no definite first character. For example, (a*|b*)[cd] would
2515 previously give no result from pcre_study(). Now it recognizes that the
2516 first character must be a, b, c, or d.
2518 21. There was an incorrect error "recursive call could loop indefinitely" if
2519 a subpattern (or the entire pattern) that was being tested for matching an
2520 empty string contained only one non-empty item after a nested subpattern.
2521 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2522 incorrectly, because the \d was being skipped in the check.
2524 22. The pcretest program now has a new pattern option /B and a command line
2525 option -b, which is equivalent to adding /B to every pattern. This causes
2526 it to show the compiled bytecode, without the additional information that
2527 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2528 is the same as /B/I).
2530 23. A new optimization is now able automatically to treat some sequences such
2531 as a*b as a*+b. More specifically, if something simple (such as a character
2532 or a simple class like \d) has an unlimited quantifier, and is followed by
2533 something that cannot possibly match the quantified thing, the quantifier
2534 is automatically "possessified".
2536 24. A recursive reference to a subpattern whose number was greater than 39
2537 went wrong under certain circumstances in UTF-8 mode. This bug could also
2538 have affected the operation of pcre_study().
2540 25. Realized that a little bit of performance could be had by replacing
2541 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2543 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2545 27. Possessive quantifiers such as a++ were previously implemented by turning
2546 them into atomic groups such as ($>a+). Now they have their own opcodes,
2547 which improves performance. This includes the automatically created ones
2548 from 23 above.
2550 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2551 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2552 the first matched character to be a colon. This applied both to named and
2553 numbered groups.
2555 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2557 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2558 building PCRE on Windows easier, so I have included it in the distribution.
2560 31. There is now a check in pcretest against a ridiculously large number being
2561 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2562 loop, the loop is abandoned.
2564 32. Forward references to subpatterns in conditions such as (?(2)...) where
2565 subpattern 2 is defined later cause pcre_compile() to search forwards in
2566 the pattern for the relevant set of parentheses. This search went wrong
2567 when there were unescaped parentheses in a character class, parentheses
2568 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2570 33. "Subroutine" calls and backreferences were previously restricted to
2571 referencing subpatterns earlier in the regex. This restriction has now
2572 been removed.
2574 34. Added a number of extra features that are going to be in Perl 5.10. On the
2575 whole, these are just syntactic alternatives for features that PCRE had
2576 previously implemented using the Python syntax or my own invention. The
2577 other formats are all retained for compatibility.
2579 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2580 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2581 also .NET compatible.
2583 (b) A recursion or subroutine call to a named group can now be defined as
2584 (?&name) as well as (?P>name).
2586 (c) A backreference to a named group can now be defined as \k<name> or
2587 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2588 5.10, are also .NET compatible.
2590 (d) A conditional reference to a named group can now use the syntax
2591 (?(<name>) or (?('name') as well as (?(name).
2593 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2594 groups (named and numbered) that are never evaluated inline, but can be
2595 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2596 is always false. There may be only one alternative in such a group.
2598 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2599 as the simple (?(R). The condition is true only if the most recent
2600 recursion is that of the given number or name. It does not search out
2601 through the entire recursion stack.
2603 (g) The escape \gN or \g{N} has been added, where N is a positive or
2604 negative number, specifying an absolute or relative reference.
2606 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2607 some "unreachable code" warnings.
2609 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2610 things, this adds five new scripts.
2612 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2613 There were also incompatibilities regarding the handling of \Q..\E inside
2614 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2615 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2617 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2618 matches an empty string, and forcibly breaks the loop. There were bugs in
2619 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2620 against aaaa the result was just "a" rather than "aaaa", for example. Two
2621 separate and independent bugs (that affected different cases) have been
2622 fixed.
2624 39. Refactored the code to abolish the use of different opcodes for small
2625 capturing bracket numbers. This is a tidy that I avoided doing when I
2626 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2627 The new approach is not only tidier, it makes it possible to reduce the
2628 memory needed to fix the previous bug (38).
2630 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2631 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2632 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2633 mode.
2635 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2636 report.
2638 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2639 copy construction and assignment in the C++ wrapper.
2641 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2642 couple of bugs that could have given wrong results in the "--newline=crlf"
2643 case.
2645 44. Added a number of casts and did some reorganization of signed/unsigned int
2646 variables following suggestions from Dair Grant. Also renamed the variable
2647 "this" as "item" because it is a C++ keyword.
2649 45. Arranged for dftables to add
2651 #include "pcre_internal.h"
2653 to pcre_chartables.c because without it, gcc 4.x may remove the array
2654 definition from the final binary if PCRE is built into a static library and
2655 dead code stripping is activated.
2657 46. For an unanchored pattern, if a match attempt fails at the start of a
2658 newline sequence, and the newline setting is CRLF or ANY, and the next two
2659 characters are CRLF, advance by two characters instead of one.
2662 Version 6.7 04-Jul-06
2663 ---------------------
2665 1. In order to handle tests when input lines are enormously long, pcretest has
2666 been re-factored so that it automatically extends its buffers when
2667 necessary. The code is crude, but this _is_ just a test program. The
2668 default size has been increased from 32K to 50K.
2670 2. The code in pcre_study() was using the value of the re argument before
2671 testing it for NULL. (Of course, in any sensible call of the function, it
2672 won't be NULL.)
2674 3. The memmove() emulation function in pcre_internal.h, which is used on
2675 systems that lack both memmove() and bcopy() - that is, hardly ever -
2676 was missing a "static" storage class specifier.
2678 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2679 containing an extended class (one that cannot be represented by a bitmap
2680 because it contains high-valued characters or Unicode property items, e.g.
2681 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2682 pattern, but PCRE should not loop if you do not (it no longer does).
2683 [Detail: two cases were found: (a) a repeated subpattern containing an
2684 extended class; (b) a recursive reference to a subpattern that followed a
2685 previous extended class. It wasn't skipping over the extended class
2686 correctly when UTF-8 mode was not set.]
2688 5. A negated single-character class was not being recognized as fixed-length
2689 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2690 compile error "lookbehind assertion is not fixed length".
2692 6. The RunPerlTest auxiliary script was showing an unexpected difference
2693 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2694 write a Perl script that can interpret lines of an input file either as
2695 byte characters or as UTF-8, which is what "perltest" was being required to
2696 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2697 can't do is switch easily at run time between having the "use utf8;" pragma
2698 or not. In the end, I fudged it by using the RunPerlTest script to insert
2699 "use utf8;" explicitly for the UTF-8 tests.
2701 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2702 the end of the subject string, contrary to the documentation and to what
2703 Perl does. This was true of both matching functions. Now it matches only at
2704 the start of the subject and immediately after *internal* newlines.
2706 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2707 a pointer to an int instead of a pointer to an unsigned long int. This
2708 caused problems on 64-bit systems.
2710 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2711 instance of the 'standard' template library not being so standard".
2713 10. There was no check on the number of named subpatterns nor the maximum
2714 length of a subpattern name. The product of these values is used to compute
2715 the size of the memory block for a compiled pattern. By supplying a very
2716 long subpattern name and a large number of named subpatterns, the size
2717 computation could be caused to overflow. This is now prevented by limiting
2718 the length of names to 32 characters, and the number of named subpatterns
2719 to 10,000.
2721 11. Subpatterns that are repeated with specific counts have to be replicated in
2722 the compiled pattern. The size of memory for this was computed from the
2723 length of the subpattern and the repeat count. The latter is limited to
2724 65535, but there was no limit on the former, meaning that integer overflow
2725 could in principle occur. The compiled length of a repeated subpattern is
2726 now limited to 30,000 bytes in order to prevent this.
2728 12. Added the optional facility to have named substrings with the same name.
2730 13. Added the ability to use a named substring as a condition, using the
2731 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2732 are numbers (not recommended). Forward references are permitted.
2734 14. Added forward references in named backreferences (if you see what I mean).
2736 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2737 pattern could run off the end of the subject. For example, the pattern
2738 "(?s)(.{1,5})"8 did this with the subject "ab".
2740 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2741 PCRE_CASELESS was set when matching characters that were quantified with ?
2742 or *.
2744 17. A character class other than a single negated character that had a minimum
2745 but no maximum quantifier - for example [ab]{6,} - was not handled
2746 correctly by pce_dfa_exec(). It would match only one character.
2748 18. A valid (though odd) pattern that looked like a POSIX character
2749 class but used an invalid character after [ (for example [[,abc,]]) caused
2750 pcre_compile() to give the error "Failed: internal error: code overflow" or
2751 in some cases to crash with a glibc free() error. This could even happen if
2752 the pattern terminated after [[ but there just happened to be a sequence of
2753 letters, a binary zero, and a closing ] in the memory that followed.
2755 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2756 over the years. Originally (before any Unicode support), just the bottom 8
2757 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2758 output from "man perlunicode" includes this:
2760 The regular expression compiler produces polymorphic opcodes. That
2761 is, the pattern adapts to the data and automatically switches to
2762 the Unicode character scheme when presented with Unicode data--or
2763 instead uses a traditional byte scheme when presented with byte
2764 data.
2766 Sadly, a wide octal escape does not cause a switch, and in a string with
2767 no other multibyte characters, these octal escapes are treated as before.
2768 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2769 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2770 Unicode string.
2772 I have not perpetrated such confusion in PCRE. Up till now, it took just
2773 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2774 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2775 translate to the appropriate multibyte character.
2777 29. Applied some refactoring to reduce the number of warnings from Microsoft
2778 and Borland compilers. This has included removing the fudge introduced
2779 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2780 a warning about an unused variable.
2782 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2783 characters since release 4.0, because Perl (from release 5.004) does not.
2784 [Or at least, is documented not to: some releases seem to be in conflict
2785 with the documentation.] However, when a pattern was studied with
2786 pcre_study() and all its branches started with \s, PCRE still included VT
2787 as a possible starting character. Of course, this did no harm; it just
2788 caused an unnecessary match attempt.
2790 22. Removed a now-redundant internal flag bit that recorded the fact that case
2791 dependency changed within the pattern. This was once needed for "required
2792 byte" processing, but is no longer used. This recovers a now-scarce options
2793 bit. Also moved the least significant internal flag bit to the most-
2794 significant bit of the word, which was not previously used (hangover from
2795 the days when it was an int rather than a uint) to free up another bit for
2796 the future.
2798 23. Added support for CRLF line endings as well as CR and LF. As well as the
2799 default being selectable at build time, it can now be changed at runtime
2800 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2801 specify that it is scanning data with non-default line endings.
2803 24. Changed the definition of CXXLINK to make it agree with the definition of
2804 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2806 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2807 recursions. This makes a big different to stack usage for some patterns.
2809 26. If a subpattern containing a named recursion or subroutine reference such
2810 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2811 the space required for the compiled pattern went wrong and gave too small a
2812 value. Depending on the environment, this could lead to "Failed: internal
2813 error: code overflow at offset 49" or "glibc detected double free or
2814 corruption" errors.
2816 27. Applied patches from Google (a) to support the new newline modes and (b) to
2817 advance over multibyte UTF-8 characters in GlobalReplace.
2819 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2820 difference for some implementation of PCRE in some Windows version.
2822 29. Added some extra testing facilities to pcretest:
2824 \q<number> in a data line sets the "match limit" value
2825 \Q<number> in a data line sets the "match recursion limt" value
2826 -S <number> sets the stack size, where <number> is in megabytes
2828 The -S option isn't available for Windows.
2831 Version 6.6 06-Feb-06
2832 ---------------------
2834 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2835 in pcreposix.h. I have copied the definition from pcre.h.
2837 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2838 because pcre.h is no longer a built file.
2840 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2841 not normally included in the compiled code.
2844 Version 6.5 01-Feb-06
2845 ---------------------
2847 1. When using the partial match feature with pcre_dfa_exec(), it was not
2848 anchoring the second and subsequent partial matches at the new starting
2849 point. This could lead to incorrect results. For example, with the pattern
2850 /1234/, partially matching against "123" and then "a4" gave a match.
2852 2. Changes to pcregrep:
2854 (a) All non-match returns from pcre_exec() were being treated as failures
2855 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2856 error message is output. Some extra information is given for the
2858 probably the only errors that are likely to be caused by users (by
2859 specifying a regex that has nested indefinite repeats, for instance).
2860 If there are more than 20 of these errors, pcregrep is abandoned.
2862 (b) A binary zero was treated as data while matching, but terminated the
2863 output line if it was written out. This has been fixed: binary zeroes
2864 are now no different to any other data bytes.
2866 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2867 used to set a locale for matching. The --locale=xxxx long option has
2868 been added (no short equivalent) to specify a locale explicitly on the
2869 pcregrep command, overriding the environment variables.
2871 (d) When -B was used with -n, some line numbers in the output were one less
2872 than they should have been.
2874 (e) Added the -o (--only-matching) option.
2876 (f) If -A or -C was used with -c (count only), some lines of context were
2877 accidentally printed for the final match.
2879 (g) Added the -H (--with-filename) option.
2881 (h) The combination of options -rh failed to suppress file names for files
2882 that were found from directory arguments.
2884 (i) Added the -D (--devices) and -d (--directories) options.
2886 (j) Added the -F (--fixed-strings) option.
2888 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2890 (l) Added the --colo(u)r option.
2892 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2893 is not present by default.
2895 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2896 items such as (?R) or (?1), when the recursion could match a number of
2897 alternatives. If it matched one of the alternatives, but subsequently,
2898 outside the recursion, there was a failure, the code tried to back up into
2899 the recursion. However, because of the way PCRE is implemented, this is not
2900 possible, and the result was an incorrect result from the match.
2902 In order to prevent this happening, the specification of recursion has
2903 been changed so that all such subpatterns are automatically treated as
2904 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2906 4. I had overlooked the fact that, in some locales, there are characters for
2907 which isalpha() is true but neither isupper() nor islower() are true. In
2908 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2909 and ordfeminine) are like this. This affected the treatment of \w and \W
2910 when they appeared in character classes, but not when they appeared outside
2911 a character class. The bit map for "word" characters is now created
2912 separately from the results of isalnum() instead of just taking it from the
2913 upper, lower, and digit maps. (Plus the underscore character, of course.)
2915 5. The above bug also affected the handling of POSIX character classes such as
2916 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2917 permanent tables. Instead, the bit maps for such a class were previously
2918 created as the appropriate unions of the upper, lower, and digit bitmaps.
2919 Now they are created by subtraction from the [[:word:]] class, which has
2920 its own bitmap.
2922 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2923 It is created by subtracting the vertical space characters (\x09, \x0a,
2924 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2925 subtraction was done in the overall bitmap for a character class, meaning
2926 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2927 be recognized. This bug has been fixed.
2929 7. Patches from the folks at Google:
2931 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2932 real life, but is still worth protecting against".
2934 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2935 regular expressions".
2937 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2938 have it.
2940 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2941 "configure" and the latter not, in order to fix a problem somebody had
2942 with compiling the Arg class on HP-UX.
2944 (e) Improve the error-handling of the C++ wrapper a little bit.
2946 (f) New tests for checking recursion limiting.
2948 8. The pcre_memmove() function, which is used only if the environment does not
2949 have a standard memmove() function (and is therefore rarely compiled),
2950 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2951 returning a result (though PCRE never actually uses the result).
2953 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2954 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2955 returned instead of calling malloc() with an overflowing number that would
2956 most likely cause subsequent chaos.
2958 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2960 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2961 with this option is matched, the nmatch and pmatch options of regexec() are
2962 ignored.
2964 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2965 provided in case anyone wants to the the POSIX interface with UTF-8
2966 strings.
2968 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2969 C++ linking (needed for some HP-UX environments).
2971 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2972 (unused parameter) and in the pcre_printint() function (omitted "default"
2973 switch label when the default is to do nothing).
2975 15. Added some code to make it possible, when PCRE is compiled as a C++
2976 library, to replace subject pointers for pcre_exec() with a smart pointer
2977 class, thus making it possible to process discontinuous strings.
2979 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2980 much the same function. They were added by different people who were trying
2981 to make PCRE easy to compile on non-Unix systems. It has been suggested
2982 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2983 for compiling on Windows systems. I have therefore replaced it with
2984 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2985 defaults to "extern" for C or "extern C" for C++, which works fine on
2986 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2987 SCOPE with something explicit in config.h. In addition:
2989 (a) pcreposix.h still had just "extern" instead of either of these macros;
2990 I have replaced it with PCRE_DATA_SCOPE.
2992 (b) Functions such as _pcre_xclass(), which are internal to the library,
2993 but external in the C sense, all had PCRE_EXPORT in their definitions.
2994 This is apparently wrong for the Windows case, so I have removed it.
2995 (It makes no difference on Unix-like systems.)
2997 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2998 of recursive calls to match(). This is different to MATCH_LIMIT because
2999 that limits the total number of calls to match(), not all of which increase
3000 the depth of recursion. Limiting the recursion depth limits the amount of
3001 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3002 when PCRE is compiled, and changed at run time. A patch from Google adds
3003 this functionality to the C++ interface.
3005 18. Changes to the handling of Unicode character properties:
3007 (a) Updated the table to Unicode 4.1.0.
3009 (b) Recognize characters that are not in the table as "Cn" (undefined).
3011 (c) I revised the way the table is implemented to a much improved format
3012 which includes recognition of ranges. It now supports the ranges that
3013 are defined in UnicodeData.txt, and it also amalgamates other
3014 characters into ranges. This has reduced the number of entries in the
3015 table from around 16,000 to around 3,000, thus reducing its size
3016 considerably. I realized I did not need to use a tree structure after
3017 all - a binary chop search is just as efficient. Having reduced the
3018 number of entries, I extended their size from 6 bytes to 8 bytes to
3019 allow for more data.
3021 (d) Added support for Unicode script names via properties such as \p{Han}.
3023 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3024 matching that character.
3026 20. When matching a repeated Unicode property with a minimum greater than zero,
3027 (for example \pL{2,}), PCRE could look past the end of the subject if it
3028 reached it while seeking the minimum number of characters. This could
3029 happen only if some of the characters were more than one byte long, because
3030 there is a check for at least the minimum number of bytes.
3032 21. Refactored the implementation of \p and \P so as to be more general, to
3033 allow for more different types of property in future. This has changed the
3034 compiled form incompatibly. Anybody with saved compiled patterns that use
3035 \p or \P will have to recompile them.
3037 22. Added "Any" and "L&" to the supported property types.
3039 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3040 but give a compile time error if the value is greater than 0xff.
3042 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3043 accidentally not being installed or uninstalled.
3045 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3046 made were to insert the current release number. This seemed silly, because
3047 it made things harder for people building PCRE on systems that don't run
3048 "configure". I have turned pcre.h into a distributed file, no longer built
3049 by "configure", with the version identification directly included. There is
3050 no longer a pcre.h.in file.
3052 However, this change necessitated a change to the pcre-config script as
3053 well. It is built from pcre-config.in, and one of the substitutions was the
3054 release number. I have updated configure.ac so that ./configure now finds
3055 the release number by grepping pcre.h.
3057 26. Added the ability to run the tests under valgrind.
3060 Version 6.4 05-Sep-05
3061 ---------------------
3063 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3064 "--" to be printed when multiple files were scanned, even when none of the
3065 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3066 consider it to be a bug, and have restored the previous behaviour.
3068 2. A couple of code tidies to get rid of compiler warnings.
3070 3. The pcretest program used to cheat by referring to symbols in the library
3071 whose names begin with _pcre_. These are internal symbols that are not
3072 really supposed to be visible externally, and in some environments it is
3073 possible to suppress them. The cheating is now confined to including
3074 certain files from the library's source, which is a bit cleaner.
3076 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3077 file's purpose clearer.
3079 5. Reorganized pcre_ucp_findchar().
3082 Version 6.3 15-Aug-05
3083 ---------------------
3085 1. The file libpcre.pc.in did not have general read permission in the tarball.
3087 2. There were some problems when building without C++ support:
3089 (a) If C++ support was not built, "make install" and "make test" still
3090 tried to test it.
3092 (b) There were problems when the value of CXX was explicitly set. Some
3093 changes have been made to try to fix these, and ...
3095 (c) --disable-cpp can now be used to explicitly disable C++ support.
3097 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3098 backslash in a target when C++ was disabled. This confuses some
3099 versions of "make", apparently. Using an intermediate variable solves
3100 this. (Same for CPP_LOBJ.)
3102 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3103 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3104 necessary on certain architectures.
3106 4. Added a setting of -export-symbols-regex to the link command to remove
3107 those symbols that are exported in the C sense, but actually are local
3108 within the library, and not documented. Their names all begin with
3109 "_pcre_". This is not a perfect job, because (a) we have to except some
3110 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3111 available (and never for static libraries). I have made a note to try to
3112 find a way round (a) in the future.
3115 Version 6.2 01-Aug-05
3116 ---------------------
3118 1. There was no test for integer overflow of quantifier values. A construction
3119 such as {1111111111111111} would give undefined results. What is worse, if
3120 a minimum quantifier for a parenthesized subpattern overflowed and became
3121 negative, the calculation of the memory size went wrong. This could have
3122 led to memory overwriting.
3124 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3126 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3127 operating environments where this matters.
3129 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3130 PCRE options from within the C++ wrapper.
3132 5. Named capturing subpatterns were not being correctly counted when a pattern
3133 was compiled. This caused two problems: (a) If there were more than 100
3134 such subpatterns, the calculation of the memory needed for the whole
3135 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3136 back references of the form \12, where the number was greater than 9, were
3137 not recognized as back references, even though there were sufficient
3138 previous subpatterns.
3140 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3141 versions of gcc, e.g. 2.95.4.
3144 Version 6.1 21-Jun-05
3145 ---------------------
3147 1. There was one reference to the variable "posix" in pcretest.c that was not
3148 surrounded by "#if !defined NOPOSIX".
3150 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3151 the cross-check on the old pcre_info() function, for the benefit of the
3152 cut-down version of PCRE that is currently imported into Exim.
3154 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3155 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3156 patterns that start (?i) but I don't think that matters. The use of (?i) is
3157 just an example; this all applies to the other options as well.
3159 4. Since libtool seems to echo the compile commands it is issuing, the output
3160 from "make" can be reduced a bit by putting "@" in front of each libtool
3161 compile command.
3163 5. Patch from the folks at Google for configure.in to be a bit more thorough
3164 in checking for a suitable C++ installation before trying to compile the
3165 C++ stuff. This should fix a reported problem when a compiler was present,
3166 but no suitable headers.
3168 6. The man pages all had just "PCRE" as their title. I have changed them to
3169 be the relevant file name. I have also arranged that these names are
3170 retained in the file doc/pcre.txt, which is a concatenation in text format
3171 of all the man pages except the little individual ones for each function.
3173 7. The NON-UNIX-USE file had not been updated for the different set of source
3174 files that come with release 6. I also added a few comments about the C++
3175 wrapper.
3178 Version 6.0 07-Jun-05
3179 ---------------------
3181 1. Some minor internal re-organization to help with my DFA experiments.
3183 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3184 didn't matter for the library itself when fully configured, but did matter
3185 when compiling without UCP support, or within Exim, where the ucp files are
3186 not imported.
3188 3. Refactoring of the library code to split up the various functions into
3189 different source modules. The addition of the new DFA matching code (see
3190 below) to a single monolithic source would have made it really too
3191 unwieldy, quite apart from causing all the code to be include in a
3192 statically linked application, when only some functions are used. This is
3193 relevant even without the DFA addition now that patterns can be compiled in
3194 one application and matched in another.
3196 The downside of splitting up is that there have to be some external
3197 functions and data tables that are used internally in different modules of
3198 the library but which are not part of the API. These have all had their
3199 names changed to start with "_pcre_" so that they are unlikely to clash
3200 with other external names.
3202 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3203 a different (DFA) algorithm. Although it is slower than the original
3204 function, it does have some advantages for certain types of matching
3205 problem.
3207 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3208 including restarting after a partial match.
3210 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3211 defined when compiling for Windows was sent to me. I have put it into the
3212 code, though I have no means of testing or verifying it.
3214 7. Added the pcre_refcount() auxiliary function.
3216 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3217 match before or at the first newline in the subject string. In pcretest,
3218 the /f option on a pattern can be used to set this.
3220 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3221 would behave wrongly. This has been present in PCRE since release 4.0.
3223 10. A number of changes to the pcregrep command:
3225 (a) Refactored how -x works; insert ^(...)$ instead of setting
3226 PCRE_ANCHORED and checking the length, in preparation for adding
3227 something similar for -w.
3229 (b) Added the -w (match as a word) option.
3231 (c) Refactored the way lines are read and buffered so as to have more
3232 than one at a time available.
3234 (d) Implemented a pcregrep test script.
3236 (e) Added the -M (multiline match) option. This allows patterns to match
3237 over several lines of the subject. The buffering ensures that at least
3238 8K, or the rest of the document (whichever is the shorter) is available
3239 for matching (and similarly the previous 8K for lookbehind assertions).
3241 (f) Changed the --help output so that it now says
3243 -w, --word-regex(p)
3245 instead of two lines, one with "regex" and the other with "regexp"
3246 because that confused at least one person since the short forms are the
3247 same. (This required a bit of code, as the output is generated
3248 automatically from a table. It wasn't just a text change.)
3250 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3251 option but starts with a hyphen. Could be a pattern or a path name
3252 starting with a hyphen, for instance.
3254 (h) "-" can be given as a file name to represent stdin.
3256 (i) When file names are being printed, "(standard input)" is used for
3257 the standard input, for compatibility with GNU grep. Previously
3258 "<stdin>" was used.
3260 (j) The option --label=xxx can be used to supply a name to be used for
3261 stdin when file names are being printed. There is no short form.
3263 (k) Re-factored the options decoding logic because we are going to add
3264 two more options that take data. Such options can now be given in four
3265 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3267 (l) Added the -A, -B, and -C options for requesting that lines of context
3268 around matches be printed.
3270 (m) Added the -L option to print the names of files that do not contain
3271 any matching lines, that is, the complement of -l.
3273 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3274 continue to scan other files.
3276 (o) The -s option was incorrectly implemented. For compatibility with other
3277 greps, it now suppresses the error message for a non-existent or non-
3278 accessible file (but not the return code). There is a new option called
3279 -q that suppresses the output of matching lines, which was what -s was
3280 previously doing.
3282 (p) Added --include and --exclude options to specify files for inclusion
3283 and exclusion when recursing.
3285 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3286 Hopefully, it now does.
3288 12. Missing cast in pcre_study().
3290 13. Added an "uninstall" target to the makefile.
3292 14. Replaced "extern" in the function prototypes in Makefile.in with
3293 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3294 world, but is set differently for Windows.
3296 15. Added a second compiling function called pcre_compile2(). The only
3297 difference is that it has an extra argument, which is a pointer to an
3298 integer error code. When there is a compile-time failure, this is set
3299 non-zero, in addition to the error test pointer being set to point to an
3300 error message. The new argument may be NULL if no error number is required
3301 (but then you may as well call pcre_compile(), which is now just a
3302 wrapper). This facility is provided because some applications need a
3303 numeric error indication, but it has also enabled me to tidy up the way
3304 compile-time errors are handled in the POSIX wrapper.
3306 16. Added VPATH=.libs to the makefile; this should help when building with one
3307 prefix path and installing with another. (Or so I'm told by someone who
3308 knows more about this stuff than I do.)
3310 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3311 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3312 match everything, including newlines. This is not POSIX-compatible, but
3313 somebody wanted the feature. From pcretest it can be activated by using
3314 both the P and the s flags.
3316 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3318 19. libpcre.pc was being incorrectly installed as executable.
3320 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3321 it now also looks for '\r' so that it will work unmodified on Windows.
3323 21. Added Google's contributed C++ wrapper to the distribution.
3325 22. Added some untidy missing memory free() calls in pcretest, to keep
3326 Electric Fence happy when testing.
3330 Version 5.0 13-Sep-04
3331 ---------------------
3333 1. Internal change: literal characters are no longer packed up into items
3334 containing multiple characters in a single byte-string. Each character
3335 is now matched using a separate opcode. However, there may be more than one
3336 byte in the character in UTF-8 mode.
3338 2. The pcre_callout_block structure has two new fields: pattern_position and
3339 next_item_length. These contain the offset in the pattern to the next match
3340 item, and its length, respectively.
3342 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3343 insertion of callouts before each pattern item. Added the /C option to
3344 pcretest to make use of this.
3346 4. On the advice of a Windows user, the lines
3348 #if defined(_WIN32) || defined(WIN32)
3349 _setmode( _fileno( stdout ), 0x8000 );
3350 #endif /* defined(_WIN32) || defined(WIN32) */
3352 have been added to the source of pcretest. This apparently does useful
3353 magic in relation to line terminators.
3355 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3356 for the benefit of those environments where the "b" makes a difference.
3358 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3359 to know about it. I have put a hack into configure.in that adds in code
3360 to set GCC=yes if CC=icc. This seems to end up at a point in the
3361 generated configure script that is early enough to affect the setting of
3362 compiler options, which is what is needed, but I have no means of testing
3363 whether it really works. (The user who reported this had patched the
3364 generated configure script, which of course I cannot do.)
3366 LATER: After change 22 below (new libtool files), the configure script
3367 seems to know about icc (and also ecc). Therefore, I have commented out
3368 this hack in configure.in.
3370 7. Added support for pkg-config (2 patches were sent in).
3372 8. Negated POSIX character classes that used a combination of internal tables
3373 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3374 [[:^ascii]]. Typically, they would match almost any characters. The other
3375 POSIX classes were not broken in this way.
3377 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3378 to find the match, as PCRE was deluded into thinking that the match had to
3379 start at the start point or following a newline. The same bug applied to
3380 patterns with negative forward assertions or any backward assertions
3381 preceding ".*" at the start, unless the pattern required a fixed first
3382 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3384 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3385 starting at the last subject character, bytes beyond the end of the subject
3386 string were read.
3388 11. Renamed the variable "class" as "classbits" to make life easier for C++
3389 users. (Previously there was a macro definition, but it apparently wasn't
3390 enough.)
3392 12. Added the new field "tables" to the extra data so that tables can be passed
3393 in at exec time, or the internal tables can be re-selected. This allows
3394 a compiled regex to be saved and re-used at a later time by a different
3395 program that might have everything at different addresses.
3397 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3398 -R library as well as a -L library.
3400 14. The debugging options of pcretest (-d on the command line or D on a
3401 pattern) showed incorrect output for anything following an extended class
3402 that contained multibyte characters and which was followed by a quantifier.
3404 15. Added optional support for general category Unicode character properties
3405 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3406 support. It adds about 90K to the size of the library. The meanings of the
3407 inbuilt class escapes such as \d and \s have NOT been changed.
3409 16. Updated pcredemo.c to include calls to free() to release the memory for the
3410 compiled pattern.
3412 17. The generated file chartables.c was being created in the source directory
3413 instead of in the building directory. This caused the build to fail if the
3414 source directory was different from the building directory, and was
3415 read-only.
3417 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3418 file. No doubt somebody will tell me if they don't make sense... Also added
3419 Dan Mooney's comments about building on OpenVMS.
3421 19. Added support for partial matching via the PCRE_PARTIAL option for
3422 pcre_exec() and the \P data escape in pcretest.
3424 20. Extended pcretest with 3 new pattern features:
3426 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3427 write the compiled pattern to the file whose name is "rest-of-line".
3428 This is a straight binary dump of the data, with the saved pointer to
3429 the character tables forced to be NULL. The study data, if any, is
3430 written too. After writing, pcretest reads a new pattern.
3432 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3433 compiled pattern from the given file. There must not be any
3434 occurrences of "<" in the file name (pretty unlikely); if there are,
3435 pcretest will instead treat the initial "<" as a pattern delimiter.
3436 After reading in the pattern, pcretest goes on to read data lines as
3437 usual.
3439 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3440 and 16-bit fields in a compiled pattern, to simulate a pattern that
3441 was compiled on a host of opposite endianness.
3443 21. The pcre-exec() function can now cope with patterns that were compiled on
3444 hosts of opposite endianness, with this restriction:
3446 As for any compiled expression that is saved and used later, the tables
3447 pointer field cannot be preserved; the extra_data field in the arguments
3448 to pcre_exec() should be used to pass in a tables address if a value
3449 other than the default internal tables were used at compile time.
3451 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3452 now diagnosed as an error. Previously, most of the time, a negative number
3453 would have been treated as zero, but if in addition "ovector" was passed as
3454 NULL, a crash could occur.
3456 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3457 new versions from the libtool 1.5 distribution (the last one is a copy of
3458 a file called libtool.m4). This seems to have fixed the need to patch
3459 "configure" to support Darwin 1.3 (which I used to do). However, I still
3460 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3461 workstation).
3463 24. Changed the PCRE licence to be the more standard "BSD" licence.
3466 Version 4.5 01-Dec-03
3467 ---------------------
3469 1. There has been some re-arrangement of the code for the match() function so
3470 that it can be compiled in a version that does not call itself recursively.
3471 Instead, it keeps those local variables that need separate instances for
3472 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3473 needs to "recurse". Keeping track of where control must go is done by means
3474 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3475 hide most of the details from the main code, and operates only if
3476 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3477 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3478 operating.
3480 To make it easier for callers to provide specially tailored get/free
3481 functions for this usage, two new functions, pcre_stack_malloc, and
3482 pcre_stack_free, are used. They are always called in strict stacking order,
3483 and the size of block requested is always the same.
3485 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3486 PCRE has been compiled to use the stack or the heap for recursion. The
3487 -C option of pcretest uses this to show which version is compiled.
3489 A new data escape \S, is added to pcretest; it causes the amounts of store
3490 obtained and freed by both kinds of malloc/free at match time to be added
3491 to the output.
3493 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3494 what's available on my current Linux desktop machine.
3496 3. When matching a UTF-8 string, the test for a valid string at the start has
3497 been extended. If start_offset is not zero, PCRE now checks that it points
3498 to a byte that is the start of a UTF-8 character. If not, it returns
3499 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3500 this is necessary because there may be backward assertions in the pattern.
3501 When matching the same subject several times, it may save resources to use
3502 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3504 4. The code for checking the validity of UTF-8 strings has been tightened so
3505 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3506 containing "overlong sequences".
3508 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3509 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3510 should have been "&", but it just so happened that all the cases this let
3511 through by mistake were picked up later in the function.
3513 6. I had used a variable called "isblank" - this is a C99 function, causing
3514 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3516 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3517 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3518 so the version is shown; (c) stop "make test" echoing "./RunTest".
3520 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3522 9. The prototype for memmove() for systems that don't have it was using
3523 size_t, but the inclusion of the header that defines size_t was later. I've
3524 moved the #includes for the C headers earlier to avoid this.
3526 10. Added some adjustments to the code to make it easier to compiler on certain
3527 special systems:
3529 (a) Some "const" qualifiers were missing.
3530 (b) Added the macro EXPORT before all exported functions; by default this
3531 is defined to be empty.
3532 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3533 that it reads its output file name as an argument instead of writing
3534 to the standard output and assuming this can be redirected.
3536 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3537 class containing characters with values greater than 255, PCRE compilation
3538 went into a loop.
3540 12. A recursive reference to a subpattern that was within another subpattern
3541 that had a minimum quantifier of zero caused PCRE to crash. For example,
3542 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3543 recursion. If the recursively-called subpattern itself had a zero repeat,
3544 that was OK.
3546 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3547 buffer into which it was copied (for escape processing) was still set at
3548 1024, so long lines caused crashes.
3550 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3551 "internal error: code overflow...". This applied to any character class
3552 that was followed by a possessive quantifier.
3554 15. Modified the Makefile to add libpcre.la as a prerequisite for
3555 libpcreposix.la because I was told this is needed for a parallel build to
3556 work.
3558 16. If a pattern that contained .* following optional items at the start was
3559 studied, the wrong optimizing data was generated, leading to matching
3560 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3561 matching string must start with a or b or c. The correct conclusion for
3562 this pattern is that a match can start with any character.
3565 Version 4.4 13-Aug-03
3566 ---------------------
3568 1. In UTF-8 mode, a character class containing characters with values between
3569 127 and 255 was not handled correctly if the compiled pattern was studied.
3570 In fixing this, I have also improved the studying algorithm for such
3571 classes (slightly).
3573 2. Three internal functions had redundant arguments passed to them. Removal
3574 might give a very teeny performance improvement.
3576 3. Documentation bug: the value of the capture_top field in a callout is *one
3577 more than* the number of the hightest numbered captured substring.
3579 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3580 in incorrectly linking with a previously installed version. They now link
3581 explicitly with libpcre.la.
3583 5. configure.in no longer needs to recognize Cygwin specially.
3585 6. A problem in pcre.in for Windows platforms is fixed.
3587 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3588 pcretest, it used to include the size of the study block as part of its
3589 output. Unfortunately, the structure contains a field that has a different
3590 size on different hardware architectures. This meant that the tests that
3591 showed this size failed. As the block is currently always of a fixed size,
3592 this information isn't actually particularly useful in pcretest output, so
3593 I have just removed it.
3595 8. Three pre-processor statements accidentally did not start in column 1.
3596 Sadly, there are *still* compilers around that complain, even though
3597 standard C has not required this for well over a decade. Sigh.
3599 9. In pcretest, the code for checking callouts passed small integers in the
3600 callout_data field, which is a void * field. However, some picky compilers
3601 complained about the casts involved for this on 64-bit systems. Now
3602 pcretest passes the address of the small integer instead, which should get
3603 rid of the warnings.
3605 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3606 both compile and run time, and gives an error if an invalid UTF-8 sequence
3607 is found. There is a option for disabling this check in cases where the
3608 string is known to be correct and/or the maximum performance is wanted.
3610 11. In response to a bug report, I changed one line in Makefile.in from
3612 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3613 to
3614 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3616 to look similar to other lines, but I have no way of telling whether this
3617 is the right thing to do, as I do not use Windows. No doubt I'll get told
3618 if it's wrong...
3621 Version 4.3 21-May-03
3622 ---------------------
3624 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3625 Makefile.
3627 2. Some refactoring to improve the quality of the code:
3629 (i) The utf8_table... variables are now declared "const".
3631 (ii) The code for \cx, which used the "case flipping" table to upper case
3632 lower case letters, now just substracts 32. This is ASCII-specific,
3633 but the whole concept of \cx is ASCII-specific, so it seems
3634 reasonable.
3636 (iii) PCRE was using its character types table to recognize decimal and
3637 hexadecimal digits in the pattern. This is silly, because it handles
3638 only 0-9, a-f, and A-F, but the character types table is locale-
3639 specific, which means strange things might happen. A private
3640 table is now used for this - though it costs 256 bytes, a table is
3641 much faster than multiple explicit tests. Of course, the standard
3642 character types table is still used for matching digits in subject
3643 strings against \d.
3645 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3646 ending in _t are). So I've renamed it as ESC_tee.
3648 3. The first argument for regexec() in the POSIX wrapper should have been
3649 defined as "const".
3651 4. Changed pcretest to use malloc() for its buffers so that they can be
3652 Electric Fenced for debugging.
3654 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3655 to read one or more bytes before the start of the subject string. Often this
3656 had no effect on PCRE's behaviour, but in some circumstances it could
3657 provoke a segmentation fault.
3659 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3660 to try to read one or more bytes before the start of the subject string.
3662 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3663 UTF-8 support could misbehave in various ways if the subject string
3664 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3665 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3666 back over UTF-8 characters.)
3669 Version 4.2 14-Apr-03
3670 ---------------------
3672 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3674 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3675 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3676 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3677 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3678 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3680 Note: automatic setting of the BUILD variables is not yet working
3681 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3682 compile-time but not at link-time
3683 [LINK]: use for linking executables only
3684 make different versions for Windows and non-Windows
3685 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3686 libraries
3687 [LINK_FOR_BUILD]: new variable
3688 [OBJEXT]: use throughout
3689 [EXEEXT]: use throughout
3690 <winshared>: new target
3691 <wininstall>: new target
3692 <dftables.o>: use native compiler
3693 <dftables>: use native linker
3694 <install>: handle Windows platform correctly
3695 <clean>: ditto
3696 <check>: ditto
3697 copy DLL to top builddir before testing
3699 As part of these changes, -no-undefined was removed again. This was reported
3700 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3701 in any case.
3703 3. Some tidies to get rid of compiler warnings:
3705 . In the match_data structure, match_limit was an unsigned long int, whereas
3706 match_call_count was an int. I've made them both unsigned long ints.
3708 . In pcretest the fact that a const uschar * doesn't automatically cast to
3709 a void * provoked a warning.
3711 . Turning on some more compiler warnings threw up some "shadow" variables
3712 and a few more missing casts.
3714 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3715 option, a class that contained a single character with a value between 128
3716 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3718 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3719 option, a class that contained several characters, but with at least one
3720 whose value was between 128 and 255 caused PCRE to crash.
3723 Version 4.1 12-Mar-03
3724 ---------------------
3726 1. Compiling with gcc -pedantic found a couple of places where casts were
3727 needed, and a string in dftables.c that was longer than standard compilers are
3728 required to support.
3730 2. Compiling with Sun's compiler found a few more places where the code could
3731 be tidied up in order to avoid warnings.
3733 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3734 first of these names is deprecated in the latest Autoconf in favour of the name
3735 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3736 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3737 analogy I have changed it to CFLAGS_FOR_BUILD.
3739 4. Added -no-undefined to the linking command in the Makefile, because this is
3740 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3741 linking step for the pcreposix library.
3743 5. PCRE was failing to diagnose the case of two named groups with the same
3744 name.
3746 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3747 literal character that is needed in the subject for a match, and scans along to
3748 ensure that it is present before embarking on the full matching process. This
3749 saves time in cases of nested unlimited repeats that are never going to match.
3750 Problem: the scan can take a lot of time if the subject is very long (e.g.
3751 megabytes), thus penalizing straightforward matches. It is now done only if the
3752 amount of subject to be scanned is less than 1000 bytes.
3754 7. A lesser problem with the same optimization is that it was recording the
3755 first character of an anchored pattern as "needed", thus provoking a search
3756 right along the subject, even when the first match of the pattern was going to
3757 fail. The "needed" character is now not set for anchored patterns, unless it
3758 follows something in the pattern that is of non-fixed length. Thus, it still
3759 fulfils its original purpose of finding quick non-matches in cases of nested
3760 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3763 Version 4.0 17-Feb-03
3764 ---------------------
3766 1. If a comment in an extended regex that started immediately after a meta-item
3767 extended to the end of string, PCRE compiled incorrect data. This could lead to
3768 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3770 2. Moved to autoconf 2.53 and libtool 1.4.2.
3772 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3773 the special perltest8 script is no longer needed - all the tests can be run
3774 from a single perltest script.
3776 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3777 by \s. It has now been removed in PCRE. This means it isn't recognized as
3778 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3779 class [:space:] *does* include VT, thereby creating a mess.
3781 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3782 space and tab.
3784 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3785 its new features into the main test script, reducing the number of scripts.
3787 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3788 were backward compatible, and made the (?i) apply to the whole pattern, as if
3789 /i were given. Now it behaves more logically, and applies the option setting
3790 only to what follows. PCRE has been changed to follow suit. However, if it
3791 finds options settings right at the start of the pattern, it extracts them into
3792 the global options, as before. Thus, they show up in the info data.
3794 8. Added support for the \Q...\E escape sequence. Characters in between are
3795 treated as literals. This is slightly different from Perl in that $ and @ are
3796 also handled as literals inside the quotes. In Perl, they will cause variable
3797 interpolation. Note the following examples:
3799 Pattern PCRE matches Perl matches
3801 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3802 \Qabc\$xyz\E abc\$xyz abc\$xyz
3803 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3805 For compatibility with Perl, \Q...\E sequences are recognized inside character
3806 classes as well as outside them.
3808 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3809 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3810 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3811 signed/unsigned warnings.
3813 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3814 option for pcretest, so I've replaced it by a simple function that does just
3815 that job.
3817 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3818 "pcregrep -".
3820 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3821 Java package. This provides some syntactic sugar for simple cases of what my
3822 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3823 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3824 item, you can use this simplified notation. Note that only makes sense with
3825 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3826 greediness, whatever the setting of the PCRE_UNGREEDY option.
3828 13. A change of greediness default within a pattern was not taking effect at
3829 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3830 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3831 was abstracted outside.
3833 14. PCRE now supports the \G assertion. It is true when the current matching
3834 position is at the start point of the match. This differs from \A when the
3835 starting offset is non-zero. Used with the /g option of pcretest (or similar
3836 code), it works in the same way as it does for Perl's /g option. If all
3837 alternatives of a regex begin with \G, the expression is anchored to the start
3838 match position, and the "anchored" flag is set in the compiled expression.
3840 15. Some bugs concerning the handling of certain option changes within patterns
3841 have been fixed. These applied to options other than (?ims). For example,
3842 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3843 been the other way round. Some of this was related to change 7 above.
3845 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3846 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3847 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3848 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3850 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3851 mode. Unlike ".", it always matches newline, whatever the setting of
3852 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3853 assertions. Perl allows it, but it doesn't (in general) work because it can't
3854 calculate the length of the lookbehind. At least, that's the case for Perl
3855 5.8.0 - I've been told they are going to document that it doesn't work in
3856 future.
3858 18. Added an error diagnosis for escapes that PCRE does not support: these are
3859 \L, \l, \N, \P, \p, \U, \u, and \X.
3861 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3862 reading past the end of the pattern in cases such as /[abcd/.
3864 20. PCRE was getting more memory than necessary for patterns with classes that
3865 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3867 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3868 compiling PCRE for use with Virtual Pascal.
3870 22. Small fix to the Makefile to make it work properly if the build is done
3871 outside the source tree.
3873 23. Added a new extension: a condition to go with recursion. If a conditional
3874 subpattern starts with (?(R) the "true" branch is used if recursion has
3875 happened, whereas the "false" branch is used only at the top level.
3877 24. When there was a very long string of literal characters (over 255 bytes
3878 without UTF support, over 250 bytes with UTF support), the computation of how
3879 much memory was required could be incorrect, leading to segfaults or other
3880 strange effects.
3882 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3883 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3884 there was a subsequent back reference to those brackets. This meant that, for
3885 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3886 possible to check for precisely this case. All we can do is abandon the
3887 optimization if .* occurs inside capturing brackets when there are any back
3888 references whatsoever. (See below for a better fix that came later.)
3890 26. The handling of the optimization for finding the first character of a
3891 non-anchored pattern, and for finding a character that is required later in the
3892 match were failing in some cases. This didn't break the matching; it just
3893 failed to optimize when it could. The way this is done has been re-implemented.
3895 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3897 28. Added a new feature that provides some of the functionality that Perl
3898 provides with (?{...}). The facility is termed a "callout". The way it is done
3899 in PCRE is for the caller to provide an optional function, by setting
3900 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3901 global variable. By default it is unset, which disables all calling out. To get
3902 the function called, the regex must include (?C) at appropriate points. This
3903 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3904 This provides a means of identifying different callout points. When PCRE
3905 reaches such a point in the regex, if pcre_callout has been set, the external
3906 function is called. It is provided with data in a structure called
3907 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3908 matching continues; if it returns a non-zero value, the match at the current
3909 point fails. However, backtracking will occur if possible. [This was changed
3910 later and other features added - see item 49 below.]
3912 29. pcretest is upgraded to test the callout functionality. It provides a
3913 callout function that displays information. By default, it shows the start of
3914 the match and the current position in the text. There are some new data escapes
3915 to vary what happens:
3917 \C+ in addition, show current contents of captured substrings
3918 \C- do not supply a callout function
3919 \C!n return 1 when callout number n is reached
3920 \C!n!m return 1 when callout number n is reached for the mth time
3922 30. If pcregrep was called with the -l option and just a single file name, it
3923 output "<stdin>" if a match was found, instead of the file name.
3925 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3926 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3927 pcre_exec(). This saves a malloc/free per call. The default value of
3928 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3929 when configuring.
3931 32. The default maximum size of a compiled pattern is 64K. There have been a
3932 few cases of people hitting this limit. The code now uses macros to handle the
3933 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3934 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3935 configuring. Tests 2 and 5 work only with 2-byte links because they output
3936 debugging information about compiled patterns.
3938 33. Internal code re-arrangements:
3940 (a) Moved the debugging function for printing out a compiled regex into
3941 its own source file (printint.c) and used #include to pull it into
3942 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3943 separate copies.
3945 (b) Defined the list of op-code names for debugging as a macro in
3946 internal.h so that it is next to the definition of the opcodes.
3948 (c) Defined a table of op-code lengths for simpler skipping along compiled
3949 code. This is again a macro in internal.h so that it is next to the
3950 definition of the opcodes.
3952 34. Added support for recursive calls to individual subpatterns, along the
3953 lines of Robin Houston's patch (but implemented somewhat differently).
3955 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3956 allow it to read and process whole directories in Win32. This code was
3957 contributed by Lionel Fourquaux; it has not been tested by me.
3959 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3960 used to name a group. Names consist of alphanumerics and underscores, and must
3961 be unique. Back references use the syntax (?P=name) and recursive calls use
3962 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3963 numbers. The function pcre_fullinfo() can be used after compilation to extract
3964 a name/number map. There are three relevant calls:
3966 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3967 PCRE_INFO_NAMECOUNT yields the number of entries
3968 PCRE_INFO_NAMETABLE yields a pointer to the map.
3970 The map is a vector of fixed-size entries. The size of each entry depends on
3971 the length of the longest name used. The first two bytes of each entry are the
3972 group number, most significant byte first. There follows the corresponding
3973 name, zero terminated. The names are in alphabetical order.
3975 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3976 case instead of 255. Making it the same both with and without UTF-8 support
3977 means that the same test output works with both.
3979 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3980 calling malloc() with a zero argument.
3982 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3983 optimization. I've improved things by keeping a bitmap of backreferences with
3984 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3985 fact referenced, the optimization can be applied. It is unlikely that a
3986 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3987 the match to follow \n) will appear inside brackets with a number greater than
3988 31, but if it does, any back reference > 31 suppresses the optimization.
3990 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3991 of disabling numbered capturing parentheses. Any opening parenthesis that is
3992 not followed by ? behaves as if it were followed by ?: but named parentheses
3993 can still be used for capturing (and they will acquire numbers in the usual
3994 way).
3996 41. Redesigned the return codes from the match() function into yes/no/error so
3997 that errors can be passed back from deep inside the nested calls. A malloc
3998 failure while inside a recursive subpattern call now causes the
3999 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4001 42. It is now possible to set a limit on the number of times the match()
4002 function is called in a call to pcre_exec(). This facility makes it possible to
4003 limit the amount of recursion and backtracking, though not in a directly
4004 obvious way, because the match() function is used in a number of different
4005 circumstances. The count starts from zero for each position in the subject
4006 string (for non-anchored patterns). The default limit is, for compatibility, a
4007 large number, namely 10 000 000. You can change this in two ways:
4009 (a) When configuring PCRE before making, you can use --with-match-limit=n
4010 to set a default value for the compiled library.
4012 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4013 a different value is set. See 45 below.
4015 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4017 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4018 of things that can be changed at compile time. The first argument specifies
4019 what is wanted and the second points to where the information is to be placed.
4020 The current list of available information is:
4024 The output is an integer that is set to one if UTF-8 support is available;
4025 otherwise it is set to zero.
4029 The output is an integer that it set to the value of the code that is used for
4030 newline. It is either LF (10) or CR (13).
4034 The output is an integer that contains the number of bytes used for internal
4035 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4039 The output is an integer that contains the threshold above which the POSIX
4040 interface uses malloc() for output vectors. See item 31 above.
4044 The output is an unsigned integer that contains the default limit of the number
4045 of match() calls in a pcre_exec() execution. See 42 above.
4047 44. pcretest has been upgraded by the addition of the -C option. This causes it
4048 to extract all the available output from the new pcre_config() function, and to
4049 output it. The program then exits immediately.
4051 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4052 order to support additional features. One way would have been to define
4053 pcre_exec2() (for example) with extra arguments, but this would not have been
4054 extensible, and would also have required all calls to the original function to
4055 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4056 is used for passing in "extra" data from pcre_study().
4058 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4059 contains the following fields:
4061 flags a bitmap indicating which of the following fields are set
4062 study_data opaque data from pcre_study()
4063 match_limit a way of specifying a limit on match() calls for a specific
4064 call to pcre_exec()
4065 callout_data data for callouts (see 49 below)
4067 The flag bits are also defined in pcre.h, and are
4073 The pcre_study() function now returns one of these new pcre_extra blocks, with
4074 the actual study data pointed to by the study_data field, and the
4075 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4076 before. That is, this change is entirely upwards-compatible and requires no
4077 change to existing code.
4079 If you want to pass in additional data to pcre_exec(), you can either place it
4080 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4081 block.
4083 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4084 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4085 times with different match limits, until it finds the minimum value needed for
4086 pcre_exec() to complete. The value is then output. This can be instructive; for
4087 most simple matches the number is quite small, but for pathological cases it
4088 gets very large very quickly.
4090 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4091 returns the size of the data block pointed to by the study_data field in a
4092 pcre_extra block, that is, the value that was passed as the argument to
4093 pcre_malloc() when PCRE was getting memory in which to place the information
4094 created by pcre_study(). The fourth argument should point to a size_t variable.
4095 pcretest has been extended so that this information is shown after a successful
4096 pcre_study() call when information about the compiled regex is being displayed.
4098 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4099 because what follows is always an absolute path. (Later: it turns out that this
4100 is more than cosmetic for MinGW, because it doesn't like empty path
4101 components.)
4103 49. Some changes have been made to the callout feature (see 28 above):
4105 (i) A callout function now has three choices for what it returns:
4107 0 => success, carry on matching
4108 > 0 => failure at this point, but backtrack if possible
4109 < 0 => serious error, return this value from pcre_exec()
4111 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4112 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4113 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4114 use by callout functions. It will never be used by PCRE itself.
4116 (ii) The pcre_extra structure (see 45 above) has a void * field called
4117 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4118 pcre_callout_block structure has a field of the same name. The contents of
4119 the field passed in the pcre_extra structure are passed to the callout
4120 function in the corresponding field in the callout block. This makes it
4121 easier to use the same callout-containing regex from multiple threads. For
4122 testing, the pcretest program has a new data escape
4124 \C*n pass the number n (may be negative) as callout_data
4126 If the callout function in pcretest receives a non-zero value as
4127 callout_data, it returns that value.
4129 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4130 there were some redundant $(CFLAGS) in commands that are now specified as
4131 $(LINK), which already includes $(CFLAGS).
4133 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4134 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4135 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4136 one-byte characters throughout. Note that case-insensitive matching applies
4137 only to characters whose values are less than 256. PCRE doesn't support the
4138 notion of cases for higher-valued characters.
4140 (i) A character class whose characters are all within 0-255 is handled as
4141 a bit map, and the map is inverted for negative classes. Previously, a
4142 character > 255 always failed to match such a class; however it should
4143 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4145 (ii) A negated character class with a single character < 255 is coded as
4146 "not this character" (OP_NOT). This wasn't working properly when the test
4147 character was multibyte, either singly or repeated.
4149 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4150 mode, for example: \x{100}{2,3}.
4152 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4153 singly or repeated) now correctly test multibyte characters. However,
4154 PCRE doesn't recognize any characters with values greater than 255 as
4155 digits, spaces, or word characters. Such characters always match \D, \S,
4156 and \W, and never match \d, \s, or \w.
4158 (v) Classes may now contain characters and character ranges with values
4159 greater than 255. For example: [ab\x{100}-\x{400}].
4161 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4162 PCRE in UTF-8 mode.
4164 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4165 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4166 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4167 value.)
4169 53. The single man page has become too large. I have therefore split it up into
4170 a number of separate man pages. These also give rise to individual HTML pages;
4171 these are now put in a separate directory, and there is an index.html page that
4172 lists them all. Some hyperlinking between the pages has been installed.
4174 54. Added convenience functions for handling named capturing parentheses.
4176 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4177 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4178 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4179 are faulted.
4181 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4182 calling configure. These values are used when compiling the dftables.c program
4183 which is run to generate the source of the default character tables. They
4184 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4185 you will need to set these values.
4187 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4190 Version 3.9 02-Jan-02
4191 ---------------------
4193 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4195 2. If --disable-static was given, the building process failed when trying to
4196 build pcretest and pcregrep. (For some reason it was using libtool to compile
4197 them, which is not right, as they aren't part of the library.)
4200 Version 3.8 18-Dec-01
4201 ---------------------
4203 1. The experimental UTF-8 code was completely screwed up. It was packing the
4204 bytes in the wrong order. How dumb can you get?
4207 Version 3.7 29-Oct-01
4208 ---------------------
4210 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4211 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4212 this didn't happen under Solaris 8, where I normally test things.
4214 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4215 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4216 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4217 this fix an hour or so after the initial 3.7 release.)
4220 Version 3.6 23-Oct-01
4221 ---------------------
4223 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4224 offsets passed as NULL with zero offset count.
4226 2. The config.guess and config.sub files had not been updated when I moved to
4227 the latest autoconf.
4230 Version 3.5 15-Aug-01
4231 ---------------------
4233 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4234 had been forgotten.
4236 2. By using declared but undefined structures, we can avoid using "void"
4237 definitions in pcre.h while keeping the internal definitions of the structures
4238 private.
4240 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4241 user point of view, this means that both static and shared libraries are built
4242 by default, but this can be individually controlled. More of the work of
4243 handling this static/shared cases is now inside libtool instead of PCRE's make
4244 file.
4246 4. The pcretest utility is now installed along with pcregrep because it is
4247 useful for users (to test regexs) and by doing this, it automatically gets
4248 relinked by libtool. The documentation has been turned into a man page, so
4249 there are now .1, .txt, and .html versions in /doc.
4251 5. Upgrades to pcregrep:
4252 (i) Added long-form option names like gnu grep.
4253 (ii) Added --help to list all options with an explanatory phrase.
4254 (iii) Added -r, --recursive to recurse into sub-directories.
4255 (iv) Added -f, --file to read patterns from a file.
4257 6. pcre_exec() was referring to its "code" argument before testing that
4258 argument for NULL (and giving an error if it was NULL).
4260 7. Upgraded Makefile.in to allow for compiling in a different directory from
4261 the source directory.
4263 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4264 options bits, the pointer it was passed was to an int instead of to an unsigned
4265 long int. This mattered only on 64-bit systems.
4267 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4268 generated) instead of pcre.in, which it its source. Also made the same change
4269 in several of the .c files.
4271 10. A new release of gcc defines printf() as a macro, which broke pcretest
4272 because it had an ifdef in the middle of a string argument for printf(). Fixed
4273 by using separate calls to printf().
4275 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4276 script, to force use of CR or LF instead of \n in the source. On non-Unix
4277 systems, the value can be set in config.h.
4279 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4280 absolute limit. Changed the text of the error message to make this clear, and
4281 likewise updated the man page.
4283 13. The limit of 99 on the number of capturing subpatterns has been removed.
4284 The new limit is 65535, which I hope will not be a "real" limit.
4287 Version 3.4 22-Aug-00
4288 ---------------------
4290 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4292 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4295 Version 3.3 01-Aug-00
4296 ---------------------
4298 1. If an octal character was given, but the value was greater than \377, it
4299 was not getting masked to the least significant bits, as documented. This could
4300 lead to crashes in some systems.
4302 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4303 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4305 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4306 These just pass their arguments on to (pcre_free)(), but they are provided
4307 because some uses of PCRE bind it to non-C systems that can call its functions,
4308 but cannot call free() or pcre_free() directly.
4310 4. Add "make test" as a synonym for "make check". Corrected some comments in
4311 the Makefile.
4313 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4314 Makefile.
4316 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4317 command called pgrep for grepping around the active processes.
4319 7. Added the beginnings of support for UTF-8 character strings.
4321 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4322 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4323 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4324 out for the ar command.)
4327 Version 3.2 12-May-00
4328 ---------------------
4330 This is purely a bug fixing release.
4332 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4333 of ZA. This was just one example of several cases that could provoke this bug,
4334 which was introduced by change 9 of version 2.00. The code for breaking
4335 infinite loops after an iteration that matches an empty string was't working
4336 correctly.
4338 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4339 when matched against abbab (for example). After matching an empty string, it
4340 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4341 caused it to match further down the string than it should.
4343 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4344 was there because it doesn't seem to be needed, and it causes trouble on some
4345 systems, as it is not a Standard C header. It has been removed.
4347 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4348 were reported on the Macintosh. The changes were from
4350 while ((c = *(++ptr)) != 0 && c != '\n');
4351 to
4352 while ((c = *(++ptr)) != 0 && c != '\n') ;
4354 Totally extraordinary, but if that's what it takes...
4356 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4357 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4358 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4359 assumes the way PCRE uses memmove() (always moving upwards).
4361 6. PCRE is being used in one environment where strchr() is not available. There
4362 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4363 faster code anyway.
4366 Version 3.1 09-Feb-00
4367 ---------------------
4369 The only change in this release is the fixing of some bugs in Makefile.in for
4370 the "install" target:
4372 (1) It was failing to install pcreposix.h.
4374 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4377 Version 3.0 01-Feb-00
4378 ---------------------
4380 1. Add support for the /+ modifier to perltest (to output $` like it does in
4381 pcretest).
4383 2. Add support for the /g modifier to perltest.
4385 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4386 matches null strings.
4388 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4389 pattern. Perl treats empty patterns specially - it reuses the most recent
4390 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4391 effect.
4393 5. The POSIX interface was broken in that it was just handing over the POSIX
4394 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4395 required a bigger vector, with some working space on the end. This means that
4396 the POSIX wrapper now has to get and free some memory, and copy the results.
4398 6. Added some simple autoconf support, placing the test data and the
4399 documentation in separate directories, re-organizing some of the
4400 information files, and making it build pcre-config (a GNU standard). Also added
4401 libtool support for building PCRE as a shared library, which is now the
4402 default.
4404 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4405 09 are not valid octal constants. Single digits will be used for minor values
4406 less than 10.
4408 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4409 existing programs that set these in the POSIX interface can use PCRE without
4410 modification.
4412 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4413 return all that pcre_info() returns, plus additional data. The pcre_info()
4414 function is retained for compatibility, but is considered to be obsolete.
4416 10. Added experimental recursion feature (?R) to handle one common case that
4417 Perl 5.6 will be able to do with (?p{...}).
4419 11. Added support for POSIX character classes like [:alpha:], which Perl is
4420 adopting.
4423 Version 2.08 31-Aug-99
4424 ----------------------
4426 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4427 trying to match at the startoffset position, but instead was moving forward to
4428 the next newline as if a previous match had failed.
4430 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4431 and could get into a loop if a null string was matched other than at the start
4432 of the subject.
4434 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4435 be distinguished at compile time, and for completeness also added PCRE_DATE.
4437 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4438 in GnuWin32 environments.
4441 Version 2.07 29-Jul-99
4442 ----------------------
4444 1. The documentation is now supplied in plain text form and HTML as well as in
4445 the form of man page sources.
4447 2. C++ compilers don't like assigning (void *) values to other pointer types.
4448 In particular this affects malloc(). Although there is no problem in Standard
4449 C, I've put in casts to keep C++ compilers happy.
4451 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4452 should be (const char *).
4454 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4455 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4456 However, I haven't made this a standard facility. The documentation doesn't
4457 mention it, and the Makefile doesn't support it.
4459 5. The Makefile now contains an "install" target, with editable destinations at
4460 the top of the file. The pcretest program is not installed.
4462 6. pgrep -V now gives the PCRE version number and date.
4464 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4465 causing the entire string to be ignored, instead of just the last character.
4467 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4468 non-matching string, it can take a very, very long time, even for strings of
4469 quite modest length, because of the nested recursion. PCRE now does better in
4470 some of these cases. It does this by remembering the last required literal
4471 character in the pattern, and pre-searching the subject to ensure it is present
4472 before running the real match. In other words, it applies a heuristic to detect
4473 some types of certain failure quickly, and in the above example, if presented
4474 with a string that has no trailing " it gives "no match" very quickly.
4476 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4477 other alternatives are tried instead.
4480 Version 2.06 09-Jun-99
4481 ----------------------
4483 1. Change pcretest's output for amount of store used to show just the code
4484 space, because the remainder (the data block) varies in size between 32-bit and
4485 64-bit systems.
4487 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4488 start matching at. This allows lookbehinds to work when searching for multiple
4489 occurrences in a string.
4491 3. Added additional options to pcretest for testing multiple occurrences:
4493 /+ outputs the rest of the string that follows a match
4494 /g loops for multiple occurrences, using the new startoffset argument
4495 /G loops for multiple occurrences by passing an incremented pointer
4497 4. PCRE wasn't doing the "first character" optimization for patterns starting
4498 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4499 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4500 the letter 'x'. On long subject strings, this gives a significant speed-up.
4503 Version 2.05 21-Apr-99
4504 ----------------------
4506 1. Changed the type of magic_number from int to long int so that it works
4507 properly on 16-bit systems.
4509 2. Fixed a bug which caused patterns starting with .* not to work correctly
4510 when the subject string contained newline characters. PCRE was assuming
4511 anchoring for such patterns in all cases, which is not correct because .* will
4512 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4513 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4514 must be retried after every newline in the subject.
4517 Version 2.04 18-Feb-99
4518 ----------------------
4520 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4521 computation of the store needed to hold the pattern was incorrect (too large).
4522 If such patterns were nested a few deep, this could multiply and become a real
4523 problem.
4525 2. Added /M option to pcretest to show the memory requirement of a specific
4526 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4528 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4529 compiled in such a way that the backtracking after subsequent failure was
4530 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4531 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4534 Version 2.03 02-Feb-99
4535 ----------------------
4537 1. Fixed typo and small mistake in man page.
4539 2. Added 4th condition (GPL supersedes if conflict) and created separate
4540 LICENCE file containing the conditions.
4542 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4543 Perl, that is the internal \ allows the delimiter to be included in the
4544 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4545 the final delimiter, add \ to the end of the pattern (to test the error).
4547 4. Added the convenience functions for extracting substrings after a successful
4548 match. Updated pcretest to make it able to test these functions.
4551 Version 2.02 14-Jan-99
4552 ----------------------
4554 1. Initialized the working variables associated with each extraction so that
4555 their saving and restoring doesn't refer to uninitialized store.
4557 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4558 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4559 fix the problem.
4561 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4562 calls, and wasn't printing the correct value for compiling calls. Increased the
4563 default value of LOOPREPEAT, and the number of significant figures in the
4564 times.
4566 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4568 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4569 a building problem on Windows NT with a FAT file system.
4572 Version 2.01 21-Oct-98
4573 ----------------------
4575 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4576 to character tables built by pcre_maketables() in the current locale. If NULL
4577 is passed, the default tables are used.
4580 Version 2.00 24-Sep-98
4581 ----------------------
4583 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4584 it any more.
4586 2. Allow quantification of (?>) groups, and make it work correctly.
4588 3. The first character computation wasn't working for (?>) groups.
4590 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4591 end of the subject) and add 5.005's \z, which really does match only at the
4592 very end of the subject.
4594 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4596 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4597 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4598 localized options. All options to pcre_study() were also removed.
4600 7. Add other new features from 5.005:
4602 $(?<= positive lookbehind
4603 $(?<! negative lookbehind
4604 (?imsx-imsx) added the unsetting capability
4605 such a setting is global if at outer level; local otherwise
4606 (?imsx-imsx:) non-capturing groups with option setting
4607 (?(cond)re|re) conditional pattern matching
4609 A backreference to itself in a repeated group matches the previous
4610 captured string.
4612 8. General tidying up of studying (both automatic and via "study")
4613 consequential on the addition of new assertions.
4615 9. As in 5.005, unlimited repeated groups that could match an empty substring
4616 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4617 runtime if any iteration does actually match an empty substring.
4619 10. Include the RunTest script in the distribution.
4621 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4622 discrepancies, some of which were old and were also with respect to 5.004. They
4623 have now been fixed.
4626 Version 1.09 28-Apr-98
4627 ----------------------
4629 1. A negated single character class followed by a quantifier with a minimum
4630 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4631 program crashes, or just wrong answers. This did not apply to negated classes
4632 containing more than one character, or to minima other than one.
4635 Version 1.08 27-Mar-98
4636 ----------------------
4638 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4640 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4641 latter must appear before anything that relies on it in the pattern.
4644 Version 1.07 16-Feb-98
4645 ----------------------
4647 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4648 repeat of a potentially empty string).
4651 Version 1.06 23-Jan-98
4652 ----------------------
4654 1. Added Markus Oberhumer's little patches for C++.
4656 2. Literal strings longer than 255 characters were broken.
4659 Version 1.05 23-Dec-97
4660 ----------------------
4662 1. Negated character classes containing more than one character were failing if
4663 PCRE_CASELESS was set at run time.
4666 Version 1.04 19-Dec-97
4667 ----------------------
4669 1. Corrected the man page, where some "const" qualifiers had been omitted.
4671 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4672 input syntax.
4674 3. Fixed memory leak which occurred when a regex with back references was
4675 matched with an offsets vector that wasn't big enough. The temporary memory
4676 that is used in this case wasn't being freed if the match failed.
4678 4. Tidied pcretest to ensure it frees memory that it gets.
4680 5. Temporary memory was being obtained in the case where the passed offsets
4681 vector was exactly big enough.
4683 6. Corrected definition of offsetof() from change 5 below.
4685 7. I had screwed up change 6 below and broken the rules for the use of
4686 setjmp(). Now fixed.
4689 Version 1.03 18-Dec-97
4690 ----------------------
4692 1. A erroneous regex with a missing opening parenthesis was correctly
4693 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4694 on some systems.
4696 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4697 it was reported that one broken compiler failed on the former because "code" is
4698 also an independent variable.
4700 3. The erroneous regex a[]b caused an array overrun reference.
4702 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4703 fail on data ending with that character. (It was going on too far, and checking
4704 the next character, typically a binary zero.) This was specific to the
4705 optimized code for single-character negative classes.
4707 5. Added a contributed patch from the TIN world which does the following:
4709 + Add an undef for memmove, in case the the system defines a macro for it.
4711 + Add a definition of offsetof(), in case there isn't one. (I don't know
4712 the reason behind this - offsetof() is part of the ANSI standard - but
4713 it does no harm).
4715 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4716 most of the places where whitespace preceded '#'. I have given up and
4717 allowed the remaining 2 cases to be at the margin.
4719 + Rename some variables in pcre to eliminate shadowing. This seems very
4720 pedantic, but does no harm, of course.
4722 6. Moved the call to setjmp() into its own function, to get rid of warnings
4723 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4725 7. Constructs such as \d{8,} were compiling into the equivalent of
4726 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4727 outcome, but in this particular case used more store than had been allocated,
4728 which caused the bug to be discovered because it threw up an internal error.
4730 8. The debugging code in both pcre and pcretest for outputting the compiled
4731 form of a regex was going wrong in the case of back references followed by
4732 curly-bracketed repeats.
4735 Version 1.02 12-Dec-97
4736 ----------------------
4738 1. Typos in pcre.3 and comments in the source fixed.
4740 2. Applied a contributed patch to get rid of places where it used to remove
4741 'const' from variables, and fixed some signed/unsigned and uninitialized
4742 variable warnings.
4744 3. Added the "runtest" target to Makefile.
4746 4. Set default compiler flag to -O2 rather than just -O.
4749 Version 1.01 19-Nov-97
4750 ----------------------
4752 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4753 like /([ab]*)*/, that is, for classes with more than one character in them.
4755 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4756 as /((?>a*))*/ (a PCRE_EXTRA facility).
4759 Version 1.00 18-Nov-97
4760 ----------------------
4762 1. Added compile-time macros to support systems such as SunOS4 which don't have
4763 memmove() or strerror() but have other things that can be used instead.
4765 2. Arranged that "make clean" removes the executables.
4768 Version 0.99 27-Oct-97
4769 ----------------------
4771 1. Fixed bug in code for optimizing classes with only one character. It was
4772 initializing a 32-byte map regardless, which could cause it to run off the end
4773 of the memory it had got.
4775 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4778 Version 0.98 22-Oct-97
4779 ----------------------
4781 1. Fixed bug in code for handling temporary memory usage when there are more
4782 back references than supplied space in the ovector. This could cause segfaults.
4785 Version 0.97 21-Oct-97
4786 ----------------------
4788 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4790 2. Optimized negated single characters not to use a bit map.
4792 3. Brought error texts together as macro definitions; clarified some of them;
4793 fixed one that was wrong - it said "range out of order" when it meant "invalid
4794 escape sequence".
4796 4. Changed some char * arguments to const char *.
4798 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4800 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4801 pcretest.
4804 Version 0.96 16-Oct-97
4805 ----------------------
4807 1. Added a simple "pgrep" utility to the distribution.
4809 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4810 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4811 where "ddd" means "one or more decimal digits".
4813 3. Fixed serious bug. If a pattern had a back reference, but the call to
4814 pcre_exec() didn't supply a large enough ovector to record the related
4815 identifying subpattern, the match always failed. PCRE now remembers the number
4816 of the largest back reference, and gets some temporary memory in which to save
4817 the offsets during matching if necessary, in order to ensure that
4818 backreferences always work.
4820 4. Increased the compatibility with Perl in a number of ways:
4822 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4823 to request this handling. The option can be set at compile or exec time.
4825 (b) $ matches before a terminating newline by default; an option
4826 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4827 mode). The option can be set at compile or exec time.
4829 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4830 the same as Perl's. If the decimal number it represents is less than 10
4831 or there aren't that many previous left capturing parentheses, an octal
4832 escape is read. Inside a character class, it's always an octal escape,
4833 even if it is a single digit.
4835 (d) An escaped but undefined alphabetic character is taken as a literal,
4836 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4837 escapes.
4839 (e) {0} is now permitted. (The previous item is removed from the compiled
4840 pattern).
4842 5. Changed all the names of code files so that the basic parts are no longer
4843 than 10 characters, and abolished the teeny "globals.c" file.
4845 6. Changed the handling of character classes; they are now done with a 32-byte
4846 bit map always.
4848 7. Added the -d and /D options to pcretest to make it possible to look at the
4849 internals of compilation without having to recompile pcre.
4852 Version 0.95 23-Sep-97
4853 ----------------------
4855 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4856 \x20 at the start of a run of normal characters. These were being treated as
4857 real characters, instead of the source characters being re-checked.
4860 Version 0.94 18-Sep-97
4861 ----------------------
4863 1. The functions are now thread-safe, with the caveat that the global variables
4864 containing pointers to malloc() and free() or alternative functions are the
4865 same for all threads.
4867 2. Get pcre_study() to generate a bitmap of initial characters for non-
4868 anchored patterns when this is possible, and use it if passed to pcre_exec().
4871 Version 0.93 15-Sep-97
4872 ----------------------
4874 1. /(b)|(:+)/ was computing an incorrect first character.
4876 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4877 but not actually doing anything yet.
4879 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4880 as Perl does (e.g. [-az] or [az-]).
4882 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4883 all possible positions.
4885 5. Split up into different modules to avoid including unneeded functions in a
4886 compiled binary. However, compile and exec are still in one module. The "study"
4887 function is split off.
4889 6. The character tables are now in a separate module whose source is generated
4890 by an auxiliary program - but can then be edited by hand if required. There are
4891 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4892 toupper() in the code.
4894 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4895 make them global. Abolish the function for setting them, as the caller can now
4896 set them directly.
4899 Version 0.92 11-Sep-97
4900 ----------------------
4902 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4903 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4905 2. Caseless matching was not working in character classes if the characters in
4906 the pattern were in upper case.
4908 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4910 4. Make PCRE_ANCHORED public and accept as a compile option.
4912 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4913 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4914 pass them.
4916 6. Give an error if bad option bits passed at compile or run time.
4918 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4919 pcretest to cause it to pass that flag.
4921 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4922 options, and the first character, if set.
4924 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4927 Version 0.91 10-Sep-97
4928 ----------------------
4930 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4931 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4933 2. PCRE was looping on encountering an indefinitely repeated back reference to
4934 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4935 Perl does - treats the match as successful.
4937 ****


Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

  ViewVC Help
Powered by ViewVC 1.1.5